CN109658193A - Determination method, apparatus, electronic equipment and the storage medium of system object importance - Google Patents
Determination method, apparatus, electronic equipment and the storage medium of system object importance Download PDFInfo
- Publication number
- CN109658193A CN109658193A CN201811565644.XA CN201811565644A CN109658193A CN 109658193 A CN109658193 A CN 109658193A CN 201811565644 A CN201811565644 A CN 201811565644A CN 109658193 A CN109658193 A CN 109658193A
- Authority
- CN
- China
- Prior art keywords
- system object
- importance
- multidimensional characteristic
- weight
- characteristic
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000012549 training Methods 0.000 claims abstract description 93
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 52
- 238000010606 normalization Methods 0.000 claims description 47
- 241000406668 Loxodonta cyclotis Species 0.000 claims description 4
- 230000000875 corresponding effect Effects 0.000 description 103
- 238000010586 diagram Methods 0.000 description 13
- 230000001174 ascending effect Effects 0.000 description 8
- 238000006243 chemical reaction Methods 0.000 description 8
- 230000006870 function Effects 0.000 description 8
- 230000006854 communication Effects 0.000 description 6
- 238000010801 machine learning Methods 0.000 description 6
- 238000004891 communication Methods 0.000 description 5
- 238000004590 computer program Methods 0.000 description 4
- 230000002596 correlated effect Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 235000013305 food Nutrition 0.000 description 4
- 238000011002 quantification Methods 0.000 description 4
- 230000006399 behavior Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000018199 S phase Effects 0.000 description 2
- 230000003466 anti-cipated effect Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000003066 decision tree Methods 0.000 description 2
- 230000001419 dependent effect Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 235000013399 edible fruits Nutrition 0.000 description 2
- 230000005611 electricity Effects 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000007477 logistic regression Methods 0.000 description 2
- 230000006855 networking Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 230000020509 sex determination Effects 0.000 description 2
- 235000011888 snacks Nutrition 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 239000002699 waste material Substances 0.000 description 2
- 238000013461 design Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0621—Item configuration or customization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/12—Hotels or restaurants
Landscapes
- Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Finance (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Development Economics (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Primary Health Care (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the present disclosure discloses determination method, apparatus, electronic equipment and the storage medium of a kind of system object importance.This method comprises: determining the multidimensional characteristic of system object;Based on training algorithm, the corresponding feature weight of the multidimensional characteristic of the system object is determined;According to the multidimensional characteristic and feature weight of the system object, the importance of the system object is determined.The embodiment of the present disclosure measures the importance of system object using the multidimensional characteristic of system object, increase the accuracy rate of the importance data of system object, the feature weight of multidimensional characteristic is also determined using minimal error rate training algorithm simultaneously, can effectively avoid manually set system object special efficacy weight and bring subjectivity so that the importance data of system object are more objective and accurate.
Description
Technical field
This disclosure relates to field of computer technology, and in particular to a kind of determination method, apparatus of system object importance, electricity
Sub- equipment and storage medium.
Background technique
With the development of information technology, e-commerce has been deep into every field, occurs numerous information on network
Platform, such as platform of ordering is taken out, the provider of information, product on these information platforms etc. by being on information platform
Object of uniting provides various services, uses for user's on-line customization.Based on the difference of service provided by system object, each system
The user group that object is agglomerated is also different.And for information platform, it is most important that number of users and user characteristics
Data.Therefore, more users and user characteristic data can be brought for information platform, then system object is for information platform
For it is then more important.Therefore, the importance of system object how is determined, and then according to the importance of system object to facilitate
The long-run development of information platform be target and the technical issues of take corresponding strategy, be current information platform urgent need to resolve it
One.
Summary of the invention
The embodiment of the present disclosure provides the method, apparatus of system object importance a kind of, electronic equipment and computer-readable deposits
Storage media.
In a first aspect, providing a kind of method of system object importance in the embodiment of the present disclosure.
Specifically, the method for the system object importance, comprising:
Determine the multidimensional characteristic of system object;
Based on training algorithm, the corresponding feature weight of the multidimensional characteristic of the system object is determined;
According to the multidimensional characteristic and feature weight of the system object, the importance of the system object is determined.
With reference to first aspect, the disclosure determines that the multidimensional of system object is special in the first implementation of first aspect
Sign, comprising:
Determine the primitive character value of the multidimensional characteristic of the system object;
The primitive character value is normalized, determines the normalization characteristic value of the multidimensional characteristic.
With reference to first aspect and/or the first implementation of first aspect, determine that the normalization of the multidimensional characteristic is special
Value indicative, comprising:
In the importance correlation of the multidimensional characteristic and the system object, the normalizing of the multidimensional characteristic
Change the ratio that characteristic value is determined as the primitive character value Yu the system object number;
In the importance negative correlation of the multidimensional characteristic and the system object, the normalizing of the multidimensional characteristic
Change characteristic value and is determined as 1 ratio for subtracting the primitive character value Yu the system object number.
The first implementation with reference to first aspect, the disclosure in second of implementation of first aspect, according to
The multidimensional characteristic and feature weight of the system object, determine the importance of the system object, comprising:
The system pair is determined after being weighted according to normalization characteristic value of the feature weight to the multidimensional characteristic
The importance of elephant.
Second of implementation with reference to first aspect, the disclosure in the third implementation of first aspect, according to
The feature weight determines the importance of the system object, packet after being weighted to the normalization characteristic value of the multidimensional characteristic
It includes:
It is calculate by the following formula the importance values of the system object i:
Wherein, N indicates the characteristic dimension number of the system object i, wjIndicate j-th of feature in the multidimensional characteristic
Feature weight;f_scoreijIndicate the normalization characteristic value of j-th of feature of system object i.
With reference to first aspect, the first implementation of first aspect, first aspect second of implementation or first
The third implementation of aspect, the disclosure in the 4th kind of implementation of first aspect, be based on training algorithm, determine described in
The corresponding feature weight of the multidimensional characteristic of system object, comprising:
Obtain training sample;Wherein, the training sample include multiple system objects between any two importance
One comparison result;
It is weighed using the training sample and minimal error rate the training algorithm feature corresponding to the multidimensional characteristic
It is trained again, and then determines the corresponding feature weight of the multidimensional characteristic.
The 4th kind of implementation with reference to first aspect, the disclosure utilize in the 5th kind of implementation of first aspect
The training sample and the minimal error rate training algorithm feature weight corresponding to the multidimensional characteristic are trained, into
And determine the corresponding feature weight of the multidimensional characteristic, comprising:
For the corresponding target signature weight of target signature in the multidimensional characteristic, other in the multidimensional characteristic
In the case that the corresponding feature weight of feature is constant, determine multiple system objects in any value of target signature weight
Under candidate importance;
According in the training sample first comparison result and the candidate importance determine that the target is special
Levy weight.
The 5th kind of implementation with reference to first aspect, the disclosure in the 6th kind of implementation of first aspect, according to
First comparison result and the candidate importance in the training sample determine the target signature weight, comprising:
Determine candidate feature weight when the candidate importance of at least two system objects is equal;
Determine under the candidate feature weight the second comparison result of the candidate importance of system object two-by-two;
Determine the degree that is consistent of second comparison result, first comparison result corresponding with the training sample;
Wherein, the size of the degree that is consistent is related to the identical quantity of the second comparison result and the first comparison result;
The highest candidate feature weight of the degree that is consistent is determined as the target signature weight.
With reference to first aspect, the first implementation, second of implementation of first aspect, first party of first aspect
The 6th kind of implementation of the third implementation in face, the 5th kind of implementation of first aspect or first aspect, the disclosure
In the 7th kind of implementation of first aspect, the multidimensional characteristic of the system object includes at least user at nearest first
Between multiple purchase rate in the period, the average visitor of user be monovalent, new objective conversion ratio, the new number of users of cohesion, single user resources amount, new user
Single stock number and whether two or more in chain trade company.
With reference to first aspect, the first implementation, second of implementation of first aspect, first party of first aspect
The 6th kind of implementation of the third implementation in face, the 5th kind of implementation of first aspect or first aspect, the disclosure
In the 8th kind of implementation of first aspect, the method also includes: it is determined and is used according to the importance of the system object
The customer parameter of the system object.
The 8th kind of implementation with reference to first aspect, the disclosure in the 9th kind of implementation of first aspect, according to
The importance of the system object determines the customer parameter using the system object, comprising:
It determines each discrete value of the customer parameter, and determines each discrete value in each discrete value institute
Corresponding multiple first fractional bits in range;
Multiple system objects are ranked up according to the importance, and determine the sorting position of the system object
Corresponding second fractional bits in entire sequence;
The customer parameter for using the system object is determined as to first fractional bits pair equal with second fractional bits
The discrete value answered.
Second aspect, the embodiment of the present disclosure provide a kind of determining device of system object importance, comprising:
First determining module is configured to determine that the multidimensional characteristic of system object;
Second determining module is configured as determining the corresponding spy of the multidimensional characteristic of the system object based on training algorithm
Levy weight;
Third determining module is configured as multidimensional characteristic and feature weight according to the system object, determines the system
The importance of system object.
In conjunction with second aspect, the disclosure is in the first implementation of second aspect, the first determining module, comprising:
First determines submodule, is configured to determine that the primitive character value of the multidimensional characteristic of the system object;
Second determines submodule, is configured as that the primitive character value is normalized, and determines that the multidimensional is special
The normalization characteristic value of sign.
In conjunction with the first of second aspect and/or second aspect implementation, described second determines submodule, comprising:
First normalization submodule, is configured as being positively correlated in the importance of the multidimensional characteristic and the system object
When relationship, the normalization characteristic value of the multidimensional characteristic is determined as the ratio of the primitive character value and the system object number
Value;
Second normalization submodule, is configured as negatively correlated in the importance of the multidimensional characteristic and the system object
When relationship, the normalization characteristic value of the multidimensional characteristic is determined as 1 and subtracts the primitive character value and the system object number
Ratio.
In conjunction with the first implementation of second aspect, the disclosure is in second of implementation of second aspect, third
Determining module, comprising:
Third determines submodule, be configured as according to the feature weight to the normalization characteristic value of the multidimensional characteristic into
The importance of the system object is determined after row weighting.
In conjunction with second of implementation of second aspect, the disclosure is in the third implementation of second aspect, third
Determine submodule, comprising: computational submodule is configured as being calculate by the following formula the importance values of the system object i:
Wherein, N indicates the characteristic dimension number of the system object i, wjIndicate j-th of feature in the multidimensional characteristic
Feature weight;f_scoreijIndicate the normalization characteristic value of j-th of feature of system object i.
In conjunction with the first implementation of second aspect, second aspect, second of implementation or second of second aspect
The third implementation of aspect, the disclosure is in the 4th kind of implementation of second aspect, the second determining module, comprising:
First acquisition submodule is configured as obtaining training sample;Wherein, the training sample includes multiple systems
First comparison result of object importance between any two;
4th determines submodule, is configured as using the training sample and minimal error rate training algorithm to described more
The corresponding feature weight of dimensional feature is trained, and then determines the corresponding feature weight of the multidimensional characteristic.
In conjunction with the 4th kind of implementation of second aspect, the disclosure is in the 5th kind of implementation of second aspect, and the 4th
Determine submodule, comprising:
5th determines submodule, is configured as the corresponding target signature power of target signature in the multidimensional characteristic
Weight, in the case that the corresponding feature weight of other features in the multidimensional characteristic is constant, determines multiple system objects
Candidate importance under any value of target signature weight;
6th determine submodule, be configured as according in the training sample first comparison result and the time
Importance is selected to determine the target signature weight.
In conjunction with the 4th kind of implementation of second aspect, the disclosure is in the 6th kind of implementation of second aspect, and the 6th
Determine submodule, comprising:
7th determines submodule, is configured to determine that time when the candidate importance of at least two system objects is equal
Select feature weight;
8th determines submodule, is configured to determine that under the candidate feature weight the candidate importance of system object two-by-two
The second comparison result;
9th determines submodule, is configured to determine that second comparison result is corresponding with the training sample described
The degree that is consistent of first comparison result;Wherein, the size of the degree that is consistent and the second comparison result and the first comparison result
Identical quantity is related;
Tenth determines submodule, is configured as the highest candidate feature weight of the degree that is consistent being determined as the target
Feature weight.
The first implementation, second of implementation of second aspect, second party in conjunction with second aspect, second aspect
The 6th kind of implementation of the third implementation in face, the 5th kind of implementation of second aspect or second aspect, the disclosure
In the 7th kind of implementation of second aspect, the multidimensional characteristic of the system object includes at least user at nearest first
Between multiple purchase rate in the period, the average visitor of user be monovalent, new objective conversion ratio, the new number of users of cohesion, single user resources amount, new user
Single stock number and whether two or more in chain trade company.
The first implementation, second of implementation of second aspect, second party in conjunction with second aspect, second aspect
The 6th kind of implementation of the third implementation in face, the 5th kind of implementation of second aspect or second aspect, the disclosure
In the 8th kind of implementation of second aspect, described device further include:
4th determining module is configured as determining the use using the system object according to the importance of the system object
Family parameter.
In conjunction with the 8th kind of implementation of second aspect, the disclosure is described in the 9th kind of implementation of second aspect
4th determining module, comprising:
11st determines submodule, is configured to determine that each discrete value of the customer parameter, and determines described each
Discrete value corresponding multiple first fractional bits in each discrete value location;
12nd determines submodule, is configured as multiple system objects being ranked up according to the importance, and
Determine the sorting position of the system object corresponding second fractional bits in entire sequence;
13rd determines submodule, is configured as to be determined as and described second using the customer parameter of the system object
The corresponding discrete value of the first equal fractional bits of fractional bits.
The function can also execute corresponding software realization by hardware realization by hardware.The hardware or
Software includes one or more modules corresponding with above-mentioned function.
It include memory and processing in the structure of the determining device of system object importance in a possible design
Device, the determining device that the memory is used to store one or more support system object importance execute in above-mentioned first aspect
The computer instruction of the determination method of system object importance, the processor is configured to being deposited in the memory for executing
The computer instruction of storage.The determining device of the system object importance can also include communication interface, be used for system object weight
The determining device and other equipment or communication for the property wanted.
The third aspect, the embodiment of the present disclosure provide a kind of electronic equipment, including memory and processor;Wherein, described
Memory is for storing one or more computer instruction, wherein one or more computer instruction is by the processor
It executes to realize method and step described in first aspect.
Fourth aspect, the embodiment of the present disclosure provide a kind of computer readable storage medium, are used for storage system object weight
Computer instruction used in the determining device for the property wanted, it includes for executing computer involved in method in above-mentioned first aspect
Instruction.
After the embodiment of the present disclosure has determined the multidimensional characteristic of system object, system object is further determined that using training algorithm
The corresponding feature weight of multidimensional characteristic;And the important of system object is determined according to the multidimensional characteristic of system object and feature weight
Property.The technical solution measures the importance of system object using the multidimensional characteristic of system object, increases the important of system object
The accuracy rate of property data, while the embodiment of the present disclosure also uses training algorithm to determine the feature weight of multidimensional characteristic, it can be effective
Avoid manually set system object special efficacy weight and bring subjectivity so that the importance data of system object are more objective
And it is accurate.
It should be understood that above general description and following detailed description be only it is exemplary and explanatory, not
The disclosure can be limited.
Detailed description of the invention
In conjunction with attached drawing, by the detailed description of following non-limiting embodiment, the other feature of the disclosure, purpose and excellent
Point will be apparent.In the accompanying drawings:
Fig. 1 shows the flow chart of the determination method of the system object importance according to one embodiment of the disclosure;
Fig. 2 shows the flow charts of the step S102 of embodiment according to Fig. 1;
Fig. 3 shows the flow chart of the step S202 of embodiment according to Fig.2,;
Fig. 4 shows the importance values according to 5 system objects of one embodiment of the disclosure about target signature weight
Coordinate schematic diagram;
Fig. 5 shows the flow chart of the step S302 of embodiment according to Fig.3,;
Fig. 6 shows the structural block diagram of the determining device of the system object importance according to one embodiment of the disclosure;
Fig. 7 shows the structural block diagram of the second determining module 602 of embodiment according to Fig.6,;
Fig. 8 shows the structural block diagram of the 4th determining submodule 702 of embodiment according to Fig.7,;
Fig. 9 shows the structural block diagram of the 6th determining submodule 802 of embodiment according to Fig.8,;
Figure 10 is adapted for the electricity for realizing the determination method of the system object importance according to one embodiment of the disclosure
The structural schematic diagram of sub- equipment.
Specific embodiment
Hereinafter, the illustrative embodiments of the disclosure will be described in detail with reference to the attached drawings, so that those skilled in the art can
Easily realize them.In addition, for the sake of clarity, the portion unrelated with description illustrative embodiments is omitted in the accompanying drawings
Point.
In the disclosure, it should be appreciated that the term of " comprising " or " having " etc. is intended to refer to disclosed in this specification
Feature, number, step, behavior, the presence of component, part or combinations thereof, and be not intended to exclude other one or more features,
A possibility that number, step, behavior, component, part or combinations thereof exist or are added.
It also should be noted that in the absence of conflict, the feature in embodiment and embodiment in the disclosure
It can be combined with each other.The disclosure is described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.
The problem of based on prior art, the embodiment of the present disclosure propose a kind of determination side of system object importance
Method, the determination method of the system object importance comprise determining that the multidimensional characteristic of system object;Based on training algorithm, institute is determined
State the corresponding feature weight of multidimensional characteristic of system object;According to the multidimensional characteristic and feature weight of the system object, determine
The importance of the system object.
Fig. 1 shows the flow chart of the determination method of the system object importance according to one embodiment of the disclosure.Such as Fig. 1 institute
Show, the determination method of the system object importance includes the following steps S101-S103:
In step s101, the multidimensional characteristic of system object is determined;
In step s 102, it is based on training algorithm, determines the corresponding feature weight of the multidimensional characteristic of the system object;
In step s 103, according to the multidimensional characteristic and feature weight of the system object, the system object is determined
Importance.
Step S101, S102 and S103 will be hereinafter further described respectively.
Step S101
Information platform is as informant by service publication provided by the service providers such as multiple information, product mutual
In networking, for the service of each service provider offer of user's on-line customization.Information platform is virtual by different service providers
For the system object in information platform, each system object can provide one or more services again on information platform.Such as
Taking out platform of ordering virtually is each virtual store taken out in platform of ordering by various food and drink providers, and user logs in outer attraction
It eats after platform, can be ordered food into virtual store.Information platform in the embodiment of the present disclosure is not limited to above-mentioned take-away and orders
Platform, as long as user oriented, and be capable of providing one or more and possess service product and service product can be provided for user
System object platform, all within the scope of the information platform of the embodiment of the present disclosure.
System object can be it is being issued in information platform, possess the virtual right of one or more user oriented services
As, and each system object can correspond to real-life one or more service providers.One in information platform is
System object can provide one or more services for user, and user can be by provided by custom-built system object on information platform
Service, and after user has customized service by information platform, the corresponding service provider of system object can mention for user
For real-life service.Service can include but is not limited to information, products in kind, virtual product etc..
Difference, the difference of customer group, the difference of self attributes of system object based on provided service etc. has not
Same feature.The embodiment of the present disclosure extracts different characteristic possessed by different system objects, and is determined by the feature of various dimensions
Importance of the system object in information platform.User's phase that importance of the system object in information platform is agglomerated with it
It closes, such as some system object can bring more users for information platform, then the importance of the system object is higher, if
Some system object can not bring user for information platform, then system object importance for information platform is lower.It is different
The information platform of property can be different for the definition of the importance of system object, and same information platform can also be based on
Different purposes is that system object defines different importance.Regardless of which kind of situation, weight of the system object for information platform
The property wanted usually all is embodied in the different characteristic of system object, such as entity businessman is had in the corresponding actual life of system object
Some features, the system object for user provide reflected some features in service process after issuing on information platform
Deng, which feature can be specifically related to, it is related for actual conditions such as the desired orientations of system object with information platform, it does not do herein
Limitation.
It is sold other than below for ordering platform, system object multidimensional involved by the importance of cohesion client's this respect
Feature can be as shown in the table:
It will be understood by those skilled in the art that the feature in list above is only to take out the example for platform of ordering, it is different
Feature corresponding to information platform, different emphasis can be different.And it will be understood by those skilled in the art that according to the disclosure
The importance of the system object of embodiment determines that technical solution can be used for various information platforms, and be not limited to take out order it is flat
Platform.
Step S102
The feature of system object can usually quantify namely the feature of each system object can be corresponding with characteristic value,
And for different characteristic, different feature power can be determined according to influence degree of the feature for the importance of system object
Weight.The present embodiment determines the feature weight of the multidimensional characteristic of system object based on training algorithm.Feature weight can refer to system
Weight shared by every one-dimensional characteristic in the importance of object.Feature weight is mainly set by artificial mode in the prior art,
But the mode of this artificial settings will cause the judgement strong depend-ence artificial experience of system object importance, subjectivity compares
By force and it is necessary to the problems such as putting into manpower, be easy to causeing the waste of human resources.Therefore, in the present embodiment, using training algorithm
Determine the corresponding feature weight of the multidimensional characteristic of the system object, it on the one hand can be to avoid engineer's weight, on the other hand
It is more scientific objective dependent on the technical data and History Order data of system object.
Step S103
It in the present embodiment, can be according to the multidimensional characteristic of the system object in the importance for determining system object
And feature weight makes comprehensive descision, it is since feature weight is obtained based on training algorithm, more objective, therefore calculate
The importance of the system object arrived is also more objective, can really reflect importance of the system object in information platform.System
The importance of object can be presented as scoring, assessment result or evaluation result to system object etc., scoring, assessment result or comment
Valence result etc. is higher, and system object is more important.The specific algorithm of the importance of system object can be decided according to the actual requirements,
Such as can be obtained using simple linear operation, also can use machine learning algorithm and obtain, machine learning algorithm include but
It is not limited to: neural network, logistic regression, support vector machines, decision tree and Markovian decision process scheduling algorithm.
After the embodiment of the present disclosure has determined the multidimensional characteristic of system object, system object is further determined that using training algorithm
The corresponding feature weight of multidimensional characteristic;And the important of system object is determined according to the multidimensional characteristic of system object and feature weight
Property.The technical solution measures the importance of system object using the multidimensional characteristic of system object, increases the important of system object
The accuracy rate of property data, while the embodiment of the present disclosure also uses training algorithm to determine the feature weight of multidimensional characteristic, it can be effective
Avoid manually set system object special efficacy weight and bring subjectivity so that the importance data of system object are more objective
And it is accurate.
In an optional implementation of the present embodiment, the step S101 determines the multidimensional characteristic of system object
The step of, it further includes steps of
Determine the primitive character value of the multidimensional characteristic of the system object;
The primitive character value is normalized, determines the normalization characteristic value of the multidimensional characteristic.
In the optional implementation, by count or calculate etc. modes obtain the system object the multidimensional it is special
The primitive character value of sign.The quantification manner of the primitive character value of each feature is different based on the difference of feature, but substantially may be used
To be determined based on the relevant historical data of system object and/or empirical data.Such as whether agglomerate new number of users and system object
Corresponding chain (KA) trade company can be obtained by statistical, and user's (can only consider new user in some embodiments) turns
Rate can be obtained by calculation.Since the characteristic value quantification manner of every dimensional feature of system object is inconsistent,
Before the characteristic value COMPREHENSIVE CALCULATING of multidimensional characteristic, the primitive character value can be normalized, and be based on normalizing
Characteristic value after change determines the importance of system object.The characteristic value normalization method of different system object features can be different.
For example, multiple purchase rate of user's (can only consider new user in some embodiments) within nearest period first time, user
(can only consider new user in some embodiments) average visitor's unit price, user (can only consider newly to use in some embodiments
Family) the higher system user of primitive characters value such as conversion ratio, the new number of users of cohesion, the corresponding importance of system user is higher, that
Characteristic value correlation after primitive character value and normalization;And the single equal stock number of single user resources amount, new user
The higher system object of equal primitive characters value, the corresponding importance of system object is lower instead, then primitive character value and normalizing
Characteristic value negative correlation after change.Although the embodiment of the present disclosure is sold in addition order platform for be illustrated, the disclosure
It is not limited to this, that is to say, that the technical solution of the embodiment of the present disclosure can be applied to various information platforms.
In an optional implementation of the present embodiment, the normalization characteristic value of the multidimensional characteristic is determined, comprising:
In the importance correlation of the multidimensional characteristic and the system object, the normalizing of the multidimensional characteristic
Change the ratio that characteristic value is determined as the primitive character value Yu the system object number;
In the importance negative correlation of the multidimensional characteristic and the system object, the normalizing of the multidimensional characteristic
Change characteristic value and is determined as 1 ratio for subtracting the primitive character value Yu the system object number.
In the optional implementation, for every one-dimensional characteristic, original spy corresponding to this feature of all system objects
Feature of the value indicative according to ascending ascending sort, after the dimensional feature normalization of each system object is calculated according to collating sequence
Value.The characteristic value calculating process of each feature normalization of system object i are as follows:
For the feature with the characteristic value correlation after normalization, normalized characteristic value can be indicated are as follows:
Wherein, f-scoreijIndicate the normalized characteristic value of feature j of system object i, f-rankijIt indicates to all systems
After object unite according to the primitive character value ascending sort of feature j, sorting position of the system object i in the list after the sequence, M
Indicate the number of all system objects.
For the feature with the characteristic value negative correlation after normalization, normalized characteristic value can be indicated are as follows:
Wherein, f_scoreijIndicate the normalized characteristic value of feature j of system object i, f_rankijIt indicates to all systems
After object unite according to the primitive character value ascending sort of feature j, sorting position of the system object i in the list after the sequence, M
Indicate the number of all system objects.
Whether for feature " chain trade company ", the list of feature values is shown as after normalization:
In an optional implementation of the present embodiment, the step S103, i.e., according to the multidimensional of the system object
Feature and feature weight, the step of determining the importance of the system object, comprising:
The system pair is determined after being weighted according to normalization characteristic value of the feature weight to the multidimensional characteristic
The importance of elephant.
The importance of system object can be determined directly by the normalization characteristic value of the multidimensional characteristic, can also be passed through and be added
It is determined again after power.In the case of not weighting, each dimensional feature is the same for the influence degree of the importance of system object.But
It is different characteristic differences all for the influence degree of the importance of system object in practical situations, therefore after use weighting
Characteristic value determine that the importance of system object is more objective.It therefore, in the present embodiment, can be according to the feature weight pair
The normalization characteristic value of the multidimensional characteristic determines the importance of the system object after being weighted.
In an optional implementation of the present embodiment, the importance of the system object i can be calculate by the following formula
Value:
Wherein, N indicates the characteristic dimension number of the system object i, wjIndicate j-th of feature in the multidimensional characteristic
Feature weight;f_scoreijIndicate the normalization characteristic value of j-th of feature of system object i.
In an optional implementation of the present embodiment, as shown in Fig. 2, the step S102, that is, determine system object
Multidimensional characteristic the step of, further comprise the steps S201-S202:
In step s 201, training sample is obtained;Wherein, the training sample include multiple system objects two-by-two it
Between importance the first comparison result;
It is corresponding to the multidimensional characteristic using the training sample and minimal error rate training algorithm in step S202
The feature weight be trained, and then determine the corresponding feature weight of the multidimensional characteristic.
Step S201
In the present embodiment, a large amount of training sample can be obtained according to historical data and/or historical experience, and passes through people
The mode of work mark, marks the first comparison result of multiple system objects importance between any two, such as system object i's is important
Property be greater than system object j importance etc..The first ratio of the importance between any two of system object in platform of ordering is sold other than hereafter
To being illustrated for result, such as " KFC should have more resources than field teacher, i.e. KFC is more important than field teacher
Property is big ", " McDonald should be fewer than the resource of KFC, i.e. McDonald is smaller than the importance of KFC " etc., can be by these institutes
The first comparison result of system object importance between any two is stated as training sample, the quantity of these training samples can be according to tool
Body business demand is determining, in the disclosure without limitation.
Step S202
In the present embodiment, it is determined that, can be using minimal error rate training algorithm to system in the case where training sample
The feature weight of object is trained, final purpose be training is obtained the feature weight of each feature utmostly
Match with the first comparison result in training sample.
The weight of the multidimensional characteristic of system object can be adjusted using minimal error rate training algorithm (mert).Most
Minor error rate training algorithm can select an optimal feature weight for each multidimensional characteristic, and the selected feature power come out
It is best suitable for the importance relativity between the importance for the system object being calculated and known system object.Example
Such as, the corresponding businessman of system object 1 is chain businessman, and the corresponding businessman of system object 2 is a very small snacks shop, that
According to historical experience it is recognised that for information platform, the importance of system object 1 is greater than system object 2.It is minimum
Error rate training algorithm is some known relativities between the importance based on each system object, is the more of system object
An optimal feature weight is selected per one-dimensional in dimensional feature, so that the system object being calculated according to these feature weights
Importance be best suitable for above-mentioned relativity.
Such as, it is first determined the initial value of the corresponding feature weight of all dimensional features, it can be using by all dimensional features pair
The method that the feature weight answered is set as preset value, wherein preset value can be some fixed numbers, such as 1;It can also pass through
The mode that machine automatically generates random number is the corresponding feature weight assignment of all dimensional features;It can also be adopted according to training sample
The initial value of the corresponding feature weight of all dimensional features is obtained with common machine learning algorithm.Due to the mark number of training sample
According to can embody to a certain extent system object between any two important ratio pair as a result, when having determined that all dimensional features are corresponding
After the initial value of feature weight, training sample can use, by minimal error rate training algorithm to the multidimensional characteristic pair
The feature weight answered is trained, so that it is determined that the corresponding feature weight of the multidimensional characteristic.
In an optional implementation of the present embodiment, as shown in figure 3, the step S202, that is, utilize the training
Sample and the minimal error rate training algorithm feature weight corresponding to the multidimensional characteristic are trained, and then determine institute
The step of stating multidimensional characteristic corresponding feature weight, further comprises the steps S301-S302:
In step S301, for the corresponding target signature weight of target signature in the multidimensional characteristic, described more
In the case that the corresponding feature weight of other features in dimensional feature is constant, determine that multiple system objects are special in the target
Levy the candidate importance under any value of weight;
In step s 302, according in the training sample first comparison result and the candidate importance it is true
The fixed target signature weight.
Step S301
In the present embodiment, minimal error rate training algorithm can use multidimensional characteristic of the mode to system object of iteration
Feature weight optimize.During the corresponding feature weight of the multidimensional characteristic to system object optimizes, every time
The corresponding feature weight of a feature of the system object can only be optimized in iteration.For example, from the system object multidimensional
The corresponding target signature weight w of a target signature p is arbitrarily chosen in featurep, to target signature weight wpIt optimizes.?
To above-mentioned target signature weight wpWhen optimizing, the corresponding feature weight of other features in multidimensional characteristic is remained unchanged.Its
In, the corresponding feature weight of other features in multidimensional characteristic is constant to be referred to, for the corresponding feature of feature of optimized mistake
Weight, feature weight are the feature weight after optimization;For the corresponding feature weight of feature that other are not optimised, feature power
Weight can be an initial value (fixed value or random number).At this point, the system object i importance values ScoreiIt can be expressed as one
A slope is f_scoreipStraight line:
As target signature weight wpWhen taking different value, the importance values of the system object i are different, can determine institute
System object i is stated in the target signature weight wpCandidate importance values namely the slope under any value are f_scoreip
Line correspondences all y values (x value be target signature weight wp)。
It can determine on information platform that multiple system objects are weighed in the target signature using above-mentioned same method
The candidate importance that important task is anticipated under value.The embodiment of the present disclosure will be with the target signature weight w of 5 system objectspFor carry out
Illustrate, as shown in figure 4, it is f_score that the importance values Score of 5 system objects can be expressed as slope respectively1p、f_
score2p、f_score3p、f_score4pAnd f_score5pStraight line, the y value of this five line correspondences then be respectively 5 systems
The candidate importance of object.
Step S302
It, can be according to the first comparison result in training sample after the candidate importance of each system object has been determined
It is determined to farthest meet the corresponding target signature power of importance of the first comparison result from these candidate importance
Weight.
In an optional implementation of the present embodiment, as shown in figure 5, the step S302, i.e., according to the training
The step of first comparison result and the candidate importance in sample determine the target signature weight, further wraps
Include following steps S501-S504:
In step S501, candidate feature power when the candidate importance of at least two system objects is equal is determined
Weight;
In step S502, determine that the second of the candidate importance of system object compares two-by-two under the candidate feature weight
As a result.
In step S503, determine that second comparison result corresponding with the training sample described first compares knot
The degree that is consistent of fruit;Wherein, the identical quantity phase of the size of the degree that is consistent and the second comparison result and the first comparison result
It closes;
In step S504, the highest candidate feature weight of the degree that is consistent is determined as the target signature weight.
In the present embodiment, the candidate importance values of each system object can be expressed as straight line, as shown in figure 4,
There is intersection point between the corresponding straight line of candidate importance values of different system objects, and only can just change at the position of intersection point
Become the importance of system object, therefore, it is first determined when the candidate importance of at least two system objects is equal (namely two
At straight-line intersection) candidate feature weight (in Fig. 4 a total of 10 intersection points, namely can determine ten candidate features power
Weight);Secondly the second comparison result of the candidate importance of system object two-by-two is determined under described candidate feature weight;Then judge
Whether the second comparison result and corresponding first comparison result are identical, and count identical quantity, for example, a certain candidate special
Levy q0 under weight, the first comparison result a total of 10, and the same quantity of the second corresponding comparison result is 9;
And at another candidate feature weight q1, the same quantity of the second comparison result corresponding with the first comparison result is 8, that
Compared to candidate feature weight q1, it is believed that at candidate feature weight q0, the second comparison result and the first comparison result
The degree that is consistent it is higher.The highest candidate feature weight of the degree that can will finally be consistent is determined as target signature weight.
In an optional implementation of the present embodiment, the multidimensional characteristic of the system object include at least user (
New user can be only considered in some embodiments) multiple purchase rate within nearest period first time, user is (in some embodiments
In can only consider new user) average visitor is monovalent, new objective conversion ratio, the new number of users of cohesion, single user resources amount, new user are single
Two or more in equal stock number and the whether chain trade company of system object.It such as may include new objective conversion ratio and cohesion
New number of users, or including agglomerating new number of users, single user resources amount, new user singly equal stock number.Realization, which makes full use of, is
The multidimensional characteristic of system object, and a few features are not limited to, to realize the important sex determination for more objectively embodying system object.
In an optional implementation of the present embodiment, the method still further comprises following steps:
The customer parameter using the system object is determined according to the importance of the system object.
In the optional implementation, the customer parameter determines method the following steps are included: according to the weight of system object
The property wanted determines the customer parameter for using system object;Wherein, the importance of the system object is according in the embodiment of the present disclosure
The determination method of system importance determines.
In the present embodiment, after system importance has been determined, customer parameter can also be determined according to system importance.
It is the resource that user provides that customer parameter, which can include but is not limited to information platform,.In some embodiments, the weight of system object
The property wanted is higher, and information platform can be more for the resource that the user that the system object is agglomerated provides, in this way being capable of promotion system
Object agglomerates more users, provides more user resources for information platform.
In an optional implementation of the present embodiment, is determined according to the importance of the system object and use the system
It the step of customer parameter of system object, further includes steps of
It determines each discrete value of the customer parameter, and determines each discrete value in each discrete value institute
Corresponding multiple first fractional bits in range;
Multiple system objects are ranked up according to the importance, and determine the sorting position of the system object
Corresponding second fractional bits in entire sequence;
The customer parameter for using the system object is determined as to first fractional bits pair equal with second fractional bits
The discrete value answered.
In the optional implementation, the determination process of customer parameter is described by following one specific application scenarios,
It should be noted that the embodiment of the present disclosure is not limited to the application scenarios.
The importance of all system objects is sorted according to sequence from small to large first;Secondly joined according to the user of configuration
The size of number interval divides the quantile of customer parameter, and the quantile divided is determined as quartile node;Root again
According to the quartile node, the system object after sequence is grouped;Situation is finally grouped according to system object, from small to large
Parameter in the corresponding parameter section of user is determined as to the parameter of user.Such as: run configuration customer parameter section be [3,
7], then parameter section size is 5, and quantile can be respectively 0.20,0.40,0.60,0.80,1.00.First by all systems
The importance of object sorts according to sequence from small to large, secondly finds 20% quartile node, 40% quartile node, 60% respectively
Quartile node and 80% quartile node, then the customer parameter value by sorting position less than the system object of 20% quartile node
It is 3, is 4 by the customer parameter value of system object that sorting position is 20% to 40%, is 40% to 60% by sorting position
The customer parameter value of system object be 5, be by the customer parameter value of system object that sorting position is 60% to 80%
6, the customer parameter value of the system object by sorting position greater than 80% is 7.
Other details in the present embodiment can be found in the description that the above-mentioned importance to system object determines embodiment of the method,
Details are not described herein.
Following is embodiment of the present disclosure, can be used for executing embodiments of the present disclosure.
Fig. 6 shows the structural block diagram of the determining device of the system object importance according to one embodiment of the disclosure, the dress
Setting being implemented in combination with as some or all of of electronic equipment by software, hardware or both.As shown in fig. 6, institute
The determining device for stating system object importance includes the first determining module 601, the second determining module 602 and third determining module
603:
First determining module 601, is configured to determine that the multidimensional characteristic of system object;
Second determining module 602 is configured as determining that the multidimensional characteristic of the system object is corresponding based on training algorithm
Feature weight;
Third determining module 603, is configured as multidimensional characteristic and feature weight according to the system object, determine described in
The importance of system object.
The first determining module 601, the second determining module 602 and third determining module 603 will hereinafter be done respectively into
The description of one step.
First determining module 601
Information platform is as informant by service publication provided by the service providers such as multiple information, product mutual
In networking, for the service of each service provider offer of user's on-line customization.Information platform is virtual by different service providers
For the system object in information platform, each system object can provide one or more services again on information platform.Such as
Taking out platform of ordering virtually is each virtual store taken out in platform of ordering by various food and drink providers, and user logs in outer attraction
It eats after platform, can be ordered food into virtual store.Information platform in the embodiment of the present disclosure is not limited to above-mentioned take-away and orders
Platform, as long as user oriented, and be capable of providing one or more and possess service product and service product can be provided for user
System object platform, all within the scope of the information platform of the embodiment of the present disclosure.
System object can be it is being issued in information platform, possess the virtual right of one or more user oriented services
As, and each system object can correspond to real-life one or more service providers.One in information platform is
System object can provide one or more services for user, and user can be by provided by custom-built system object on information platform
Service, and after user has customized service by information platform, the corresponding service provider of system object can mention for user
For real-life service.Service can include but is not limited to information, products in kind, virtual product etc..
Difference, the difference of customer group, the difference of self attributes of system object based on provided service etc. has not
Same feature.The embodiment of the present disclosure extracts different characteristic possessed by different system objects, and is determined by the feature of various dimensions
Importance of the system object in information platform.User's phase that importance of the system object in information platform is agglomerated with it
It closes, such as some system object can bring more users for information platform, then the importance of the system object is higher, if
Some system object can not bring user for information platform, then system object importance for information platform is lower.It is different
The information platform of property can be different for the definition of the importance of system object, and same information platform can also be based on
Different purposes is that system object defines different importance.Regardless of which kind of situation, weight of the system object for information platform
The property wanted usually all is embodied in the different characteristic of system object, such as entity businessman is had in the corresponding actual life of system object
Some features, the system object for user provide reflected some features in service process after issuing on information platform
Deng, which feature can be specifically related to, it is related for actual conditions such as the desired orientations of system object with information platform, it does not do herein
Limitation.
It is sold other than below for ordering platform, system object multidimensional involved by the importance of cohesion client's this respect
Feature can be as shown in the table:
It will be understood by those skilled in the art that the feature in list above is only to take out the example for platform of ordering, it is different
Feature corresponding to information platform, different emphasis can be different.And it will be understood by those skilled in the art that according to the disclosure
The importance of the system object of embodiment determines that technical solution can be used for various information platforms, and be not limited to take out order it is flat
Platform.
Second determining module 602
The feature of system object can usually quantify namely the feature of each system object can be corresponding with characteristic value,
And for different characteristic, different feature power can be determined according to influence degree of the feature for the importance of system object
Weight.The present embodiment determines the feature weight of the multidimensional characteristic of system object based on training algorithm.Feature weight can refer to system
Weight shared by every one-dimensional characteristic in the importance of object.Feature weight is mainly set by artificial mode in the prior art,
But the mode of this artificial settings will cause the judgement strong depend-ence artificial experience of system object importance, subjectivity compares
By force and it is necessary to the problems such as putting into manpower, be easy to causeing the waste of human resources.Therefore, in the present embodiment, using training algorithm
Determine the corresponding feature weight of the multidimensional characteristic of the system object, it on the one hand can be to avoid engineer's weight, on the other hand
It is more scientific objective dependent on the technical data and History Order data of system object.
Third determining module 603
It in the present embodiment, can be according to the multidimensional characteristic of the system object in the importance for determining system object
And feature weight makes comprehensive descision, it is since feature weight is obtained based on training algorithm, more objective, therefore calculate
The importance of the system object arrived is also more objective, can really reflect importance of the system object in information platform.System
The importance of object can be presented as scoring, assessment result or evaluation result to system object etc., scoring, assessment result or comment
Valence result etc. is higher, and system object is more important.The specific algorithm of the importance of system object can be decided according to the actual requirements,
Such as can be obtained using simple linear operation, also can use machine learning algorithm and obtain, machine learning algorithm include but
It is not limited to: neural network, logistic regression, support vector machines, decision tree and Markovian decision process scheduling algorithm.
After the embodiment of the present disclosure has determined the multidimensional characteristic of system object, system object is further determined that using training algorithm
The corresponding feature weight of multidimensional characteristic;And the important of system object is determined according to the multidimensional characteristic of system object and feature weight
Property.The technical solution measures the importance of system object using the multidimensional characteristic of system object, increases the important of system object
The accuracy rate of property data, while the embodiment of the present disclosure also uses training algorithm to determine the feature weight of multidimensional characteristic, it can be effective
Avoid manually set system object special efficacy weight and bring subjectivity so that the importance data of system object are more objective
And it is accurate.
In an optional implementation of the present embodiment, first determining module 601 further comprises:
First determines submodule, is configured to determine that the primitive character value of the multidimensional characteristic of the system object;
Second determines submodule, is configured as that the primitive character value is normalized, and determines that the multidimensional is special
The normalization characteristic value of sign.
In the optional implementation, by count or calculate etc. modes obtain the system object the multidimensional it is special
The primitive character value of sign.The quantification manner of the primitive character value of each feature is different based on the difference of feature, but substantially may be used
To be determined based on the relevant historical data of system object and/or empirical data.Such as whether agglomerate new number of users and system object
Corresponding chain (KA) trade company can be obtained by statistical, and user's (can only consider new user in some embodiments) turns
Rate can be obtained by calculation.Since the characteristic value quantification manner of every dimensional feature of system object is inconsistent,
Before the characteristic value COMPREHENSIVE CALCULATING of multidimensional characteristic, the primitive character value can be normalized, and be based on normalizing
Characteristic value after change determines the importance of system object.The characteristic value normalization method of different system object features can be different.
For example, multiple purchase rate of user's (can only consider new user in some embodiments) within nearest period first time, user
(can only consider new user in some embodiments) average visitor's unit price, user (can only consider newly to use in some embodiments
Family) the higher system user of primitive characters value such as conversion ratio, the new number of users of cohesion, the corresponding importance of system user is higher, that
Characteristic value correlation after primitive character value and normalization;And the single equal stock number of single user resources amount, new user
The higher system object of equal primitive characters value, the corresponding importance of system object is lower instead, then primitive character value and normalizing
Characteristic value negative correlation after change.Although the embodiment of the present disclosure is sold in addition order platform for be illustrated, the disclosure
It is not limited to this, that is to say, that the technical solution of the embodiment of the present disclosure can be applied to various information platforms.
In an optional implementation of the present embodiment, described second determines submodule, comprising:
First normalization submodule, is configured as being positively correlated in the importance of the multidimensional characteristic and the system object
When relationship, the normalization characteristic value of the multidimensional characteristic is determined as the ratio of the primitive character value and the system object number
Value;
Second normalization submodule, is configured as negatively correlated in the importance of the multidimensional characteristic and the system object
When relationship, the normalization characteristic value of the multidimensional characteristic is determined as 1 and subtracts the primitive character value and the system object number
Ratio.
In the optional implementation, for every one-dimensional characteristic, original spy corresponding to this feature of all system objects
Feature of the value indicative according to ascending ascending sort, after the dimensional feature normalization of each system object is calculated according to collating sequence
Value.The characteristic value calculating process of each feature normalization of system object i are as follows:
For the feature with the characteristic value correlation after normalization, normalized characteristic value can be indicated are as follows:
Wherein, f_scoreijIndicate the normalized characteristic value of feature j of system object i, f_rankijIt indicates to all systems
After object unite according to the primitive character value ascending sort of feature j, sorting position of the system object i in the list after the sequence, M
Indicate the number of all system objects.
For the feature with the characteristic value negative correlation after normalization, normalized characteristic value can be indicated are as follows:
Wherein, f_scoreijIndicate the normalized characteristic value of feature j of system object i, f_rankijIt indicates to all systems
After object unite according to the primitive character value ascending sort of feature j, sorting position of the system object i in the list after the sequence, M
Indicate the number of all system objects.
Whether for feature " chain trade company ", the list of feature values is shown as after normalization:
In an optional implementation of the present embodiment, the third determining module 603, comprising:
Third determines submodule, be configured as according to the feature weight to the normalization characteristic value of the multidimensional characteristic into
The importance of the system object is determined after row weighting.
The importance of system object can be determined directly by the normalization characteristic value of the multidimensional characteristic, can also be passed through and be added
It is determined again after power.In the case of not weighting, each dimensional feature is the same for the influence degree of the importance of system object.But
It is different characteristic differences all for the influence degree of the importance of system object in practical situations, therefore after use weighting
Characteristic value determine that the importance of system object is more objective.It therefore, in the present embodiment, can be according to the feature weight pair
The normalization characteristic value of the multidimensional characteristic determines the importance of the system object after being weighted.
In an optional implementation of the present embodiment, the importance of the system object i can be calculate by the following formula
Value:
Wherein, N indicates the characteristic dimension number of the system object i,wjIndicate j-th of feature in the multidimensional characteristic
Feature weight;f_scoreijIndicate the normalization characteristic value of j-th of feature of system object i.
In an optional implementation of the present embodiment, as shown in fig. 7, second determining module 602, comprising:
First acquisition submodule 701 is configured as obtaining training sample;Wherein, the training sample includes multiple described
First comparison result of system object importance between any two;
4th determines submodule 702, is configured as using the training sample and minimal error rate training algorithm to institute
It states the corresponding feature weight of multidimensional characteristic to be trained, and then determines the corresponding feature weight of the multidimensional characteristic.
First acquisition submodule 701
In the present embodiment, a large amount of training sample can be obtained according to historical data and/or historical experience, and passes through people
The mode of work mark, marks the first comparison result of multiple system objects importance between any two, such as system object i's is important
Property be greater than system object j importance etc..The first ratio of the importance between any two of system object in platform of ordering is sold other than hereafter
To being illustrated for result, such as " KFC should have more resources than field teacher, i.e. KFC is more important than field teacher
Property is big ", " McDonald should be fewer than the resource of KFC, i.e. McDonald is smaller than the importance of KFC " etc., can be by these institutes
The first comparison result of system object importance between any two is stated as training sample, the quantity of these training samples can be according to tool
Body business demand is determining, in the disclosure without limitation.
4th determines submodule 702
In the present embodiment, it is determined that, can be using minimal error rate training algorithm to system in the case where training sample
The feature weight of object is trained, final purpose be training is obtained the feature weight of each feature utmostly
Match with the first comparison result in training sample.
The weight of the multidimensional characteristic of system object can be adjusted using minimal error rate training algorithm (mert).Most
Minor error rate training algorithm can select an optimal feature weight for each multidimensional characteristic, and the selected feature power come out
It is best suitable for the importance relativity between the importance for the system object being calculated and known system object.Example
Such as, the corresponding businessman of system object 1 is chain businessman, and the corresponding businessman of system object 2 is a very small snacks shop, that
According to historical experience it is recognised that for information platform, the importance of system object 1 is greater than system object 2.It is minimum
Error rate training algorithm is some known relativities between the importance based on each system object, is the more of system object
An optimal feature weight is selected per one-dimensional in dimensional feature, so that the system object being calculated according to these feature weights
Importance be best suitable for above-mentioned relativity.
Such as, it is first determined the initial value of the corresponding feature weight of all dimensional features, it can be using by all dimensional features pair
The method that the feature weight answered is set as preset value, wherein preset value can be some fixed numbers, such as 1;It can also pass through
The mode that machine automatically generates random number is the corresponding feature weight assignment of all dimensional features;It can also be adopted according to training sample
The initial value of the corresponding feature weight of all dimensional features is obtained with common machine learning algorithm.Due to the mark number of training sample
According to can embody to a certain extent system object between any two important ratio pair as a result, when having determined that all dimensional features are corresponding
After the initial value of feature weight, training sample can use, by minimal error rate training algorithm to the multidimensional characteristic pair
The feature weight answered is trained, so that it is determined that the corresponding feature weight of the multidimensional characteristic.
In an optional implementation of the present embodiment, as shown in figure 8, the described 4th determines submodule 702, comprising:
5th determines submodule 801, is configured as the corresponding target signature of target signature in the multidimensional characteristic
Weight determines multiple systems pair in the case that the corresponding feature weight of other features in the multidimensional characteristic is constant
As the candidate importance under any value of target signature weight;
6th determine submodule 802, be configured as according in the training sample first comparison result and institute
It states candidate importance and determines the target signature weight.
5th determines submodule 801
In the present embodiment, minimal error rate training algorithm can use multidimensional characteristic of the mode to system object of iteration
Feature weight optimize.During the corresponding feature weight of the multidimensional characteristic to system object optimizes, every time
The corresponding feature weight of a feature of the system object can only be optimized in iteration.For example, from the system object multidimensional
The corresponding target signature weight w of a target signature p is arbitrarily chosen in featurep, to target signature weight wpIt optimizes.?
To above-mentioned target signature weight wpWhen optimizing, the corresponding feature weight of other features in multidimensional characteristic is remained unchanged.Its
In, the corresponding feature weight of other features in multidimensional characteristic is constant to be referred to, for the corresponding feature of feature of optimized mistake
Weight, feature weight are the feature weight after optimization;For the corresponding feature weight of feature that other are not optimised, feature power
Weight can be an initial value (fixed value or random number).At this point, the system object i importance values ScoreiIt can be expressed as one
A slope is f_scoreipStraight line:
As target signature weight wpWhen taking different value, the importance values of the system object i are different, can determine institute
System object i is stated in the target signature weight wpCandidate importance values namely the slope under any value are f_scoreip
Line correspondences all y values (x value be target signature weight wp)。
It can determine on information platform that multiple system objects are weighed in the target signature using above-mentioned same method
The candidate importance that important task is anticipated under value.The embodiment of the present disclosure will be with the target signature weight w of 5 system objectspFor carry out
Illustrate, as shown in figure 4, it is f_score that the importance values Score of 5 system objects can be expressed as slope respectively1p、f_
score2p、f_score3p、f_score4pAnd fscore5pStraight line, the y value of this five line correspondences then be respectively 5 systems pair
The candidate importance of elephant.
6th determines submodule 802
It, can be according to the first comparison result in training sample after the candidate importance of each system object has been determined
It is determined to farthest meet the corresponding target signature power of importance of the first comparison result from these candidate importance
Weight.
In an optional implementation of the present embodiment, as shown in figure 9, the described 6th determines submodule 802, comprising:
7th determines submodule 901, when being configured to determine that the candidate importance of at least two system objects is equal
Candidate feature weight;
8th determines submodule 902, is configured as under the candidate feature weight the candidate importance of system object two-by-two
The second comparison result.
9th determines submodule 903, is configured to determine that second comparison result is corresponding with the training sample
The degree that is consistent of first comparison result;Wherein, the size of the degree that is consistent and the second comparison result and first compare knot
The identical quantity of fruit is related;
Tenth determines submodule 904, is configured as the highest candidate feature weight of the degree that is consistent being determined as described
Target signature weight.
In the present embodiment, the candidate importance values of each system object can be expressed as straight line, as shown in figure 4,
There is intersection point between the corresponding straight line of candidate importance values of different system objects, and only can just change at the position of intersection point
Become the importance of system object, therefore, determines that submodule 901 determines the time of at least two system objects by the 7th first
(a total of 10 intersection points, also can candidate feature weight when selecting importance equal (namely at two straight-line intersections) in Fig. 4
Determine ten candidate feature weights);Determine that submodule 902 determines under the candidate feature weight system two-by-two secondly by the 8th
Second comparison result of the candidate importance of object;Then determine that submodule 903 judges the second comparison result and right by the 9th
Whether the first comparison result answered is identical, and counts identical quantity, for example, the q0 under a certain candidate feature weight, first
Comparison result a total of 10, and the same quantity of the second corresponding comparison result is 9;And in another candidate feature
Under weight q1, the same quantity of the second comparison result corresponding with the first comparison result is 8, then compared to candidate feature
Weight q1, it is believed that at candidate feature weight q0, the second comparison result and the degree that is consistent of the first comparison result are higher.Most
Determine that the highest candidate feature weight of degree that can will be consistent of submodule 904 is determined as target signature weight by the tenth afterwards.
In an optional implementation of the present embodiment, the multidimensional characteristic of the system object include at least user (
New user can be only considered in some embodiments) multiple purchase rate within nearest period first time, user is (in some embodiments
In can only consider new user) average visitor is monovalent, new objective conversion ratio, the new number of users of cohesion, single user resources amount, new user are single
Two or more in equal stock number and the whether chain trade company of system object.It such as may include new objective conversion ratio and cohesion
New number of users, or including agglomerating new number of users, single user resources amount, new user singly equal stock number.Realization, which makes full use of, is
The multidimensional characteristic of system object, and a few features are not limited to, to realize the important sex determination for more objectively embodying system object.
In an optional implementation of the present embodiment, the determining device of system object importance further include:
4th determining module is configured as determining the use using the system object according to the importance of the system object
Family parameter.
In the present embodiment, after system importance has been determined, customer parameter can also be determined according to system importance.
It is the resource that user provides that customer parameter, which can include but is not limited to information platform,.In some embodiments, the weight of system object
The property wanted is higher, and information platform can be more for the resource that the user that the system object is agglomerated provides, in this way being capable of promotion system
Object agglomerates more users, provides more user resources for information platform.
In an optional implementation of the present embodiment, the 4th determining module, comprising:
11st determines submodule, is configured to determine that each discrete value of the customer parameter, and determines described each
Discrete value corresponding multiple first fractional bits in each discrete value location;
12nd determines submodule, is configured as multiple system objects being ranked up according to the importance, and
Determine the sorting position of the system object corresponding second fractional bits in entire sequence;
13rd determines submodule, is configured as to be determined as and described second using the customer parameter of the system object
The corresponding discrete value of the first equal fractional bits of fractional bits.
In the optional implementation, the determination process of customer parameter is described by following one specific application scenarios,
It should be noted that the embodiment of the present disclosure is not limited to the application scenarios.
The importance of all system objects is sorted according to sequence from small to large first;Secondly joined according to the user of configuration
The size of number interval divides the quantile of customer parameter, and the quantile divided is determined as quartile node;Root again
According to the quartile node, the system object after sequence is grouped;Situation is finally grouped according to system object, from small to large
Parameter in the corresponding parameter section of user is determined as to the parameter of user.Such as: run configuration customer parameter section be [3,
7], then parameter section size is 5, and quantile can be respectively 0.20,0.40,0.60,0.80,1.00.First by all systems
The importance of object sorts according to sequence from small to large, secondly finds 20% quartile node, 40% quartile node, 60% respectively
Quartile node and 80% quartile node, then the customer parameter value by sorting position less than the system object of 20% quartile node
It is 3, is 4 by the customer parameter value of system object that sorting position is 20% to 40%, is 40% to 60% by sorting position
The customer parameter value of system object be 5, be by the customer parameter value of system object that sorting position is 60% to 80%
6, the customer parameter value of the system object by sorting position greater than 80% is 7.
Other details in the present embodiment can be found in the description that the above-mentioned importance to system object determines embodiment of the method,
Details are not described herein.
Figure 10 is adapted for the electronics for realizing the determination method of the system object importance according to disclosure embodiment
The structural schematic diagram of equipment.
As shown in Figure 10, electronic equipment 1000 includes central processing unit (CPU) 1001, can be read-only according to being stored in
Program in memory (ROM) 1002 is loaded into the journey in random access storage device (RAM) 1003 from storage section 1008
Sequence and execute the various processing in above-mentioned embodiment shown in FIG. 1.In RAM1003, it is also stored with the behaviour of electronic equipment 1000
Various programs and data needed for making.CPU1001, ROM1002 and RAM1003 are connected with each other by bus 1004.Input/defeated
(I/O) interface 1005 is also connected to bus 1004 out.
I/O interface 1005 is connected to lower component: the importation 1006 including keyboard, mouse etc.;Including such as cathode
The output par, c 1007 of ray tube (CRT), liquid crystal display (LCD) etc. and loudspeaker etc.;Storage section including hard disk etc.
1008;And the communications portion 1009 of the network interface card including LAN card, modem etc..Communications portion 1009 passes through
Communication process is executed by the network of such as internet.Driver 1010 is also connected to I/O interface 1005 as needed.It is detachable to be situated between
Matter 1011, such as disk, CD, magneto-optic disk, semiconductor memory etc. are mounted on as needed on driver 1010, so as to
In being mounted into storage section 1008 as needed from the computer program read thereon.
Particularly, according to embodiment of the present disclosure, it is soft to may be implemented as computer above with reference to Fig. 1 method described
Part program.For example, embodiment of the present disclosure includes a kind of computer program product comprising be tangibly embodied in and its readable
Computer program on medium, the computer program include the program code for executing the method for Fig. 1.In such implementation
In mode, which can be downloaded and installed from network by communications portion 1009, and/or from detachable media
1011 are mounted.
Flow chart and block diagram in attached drawing illustrate system, method and computer according to the various embodiments of the disclosure
The architecture, function and operation in the cards of program product.In this regard, each box in course diagram or block diagram can be with
A part of a module, section or code is represented, a part of the module, section or code includes one or more
Executable instruction for implementing the specified logical function.It should also be noted that in some implementations as replacements, institute in box
The function of mark can also occur in a different order than that indicated in the drawings.For example, two boxes succeedingly indicated are practical
On can be basically executed in parallel, they can also be executed in the opposite order sometimes, and this depends on the function involved.Also it wants
It is noted that the combination of each box in block diagram and or flow chart and the box in block diagram and or flow chart, Ke Yiyong
The dedicated hardware based system of defined functions or operations is executed to realize, or can be referred to specialized hardware and computer
The combination of order is realized.
Being described in unit or module involved in disclosure embodiment can be realized by way of software, can also
It is realized in a manner of through hardware.Described unit or module also can be set in the processor, these units or module
Title do not constitute the restriction to the unit or module itself under certain conditions.
As on the other hand, the disclosure additionally provides a kind of computer readable storage medium, the computer-readable storage medium
Matter can be computer readable storage medium included in device described in above embodiment;It is also possible to individualism,
Without the computer readable storage medium in supplying equipment.Computer-readable recording medium storage has one or more than one journey
Sequence, described program is used to execute by one or more than one processor is described in disclosed method.
Above description is only the preferred embodiment of the disclosure and the explanation to institute's application technology principle.Those skilled in the art
Member is it should be appreciated that invention scope involved in the disclosure, however it is not limited to technology made of the specific combination of above-mentioned technical characteristic
Scheme, while should also cover in the case where not departing from the inventive concept, it is carried out by above-mentioned technical characteristic or its equivalent feature
Any combination and the other technical solutions formed.Such as features described above has similar function with (but being not limited to) disclosed in the disclosure
Can technical characteristic replaced mutually and the technical solution that is formed.
Claims (10)
1. a kind of determination method of system object importance characterized by comprising
Determine the multidimensional characteristic of system object;
The corresponding feature weight of multidimensional characteristic of the system object is determined based on training algorithm;
According to the multidimensional characteristic and feature weight of the system object, the importance of the system object is determined.
2. the method according to claim 1, wherein determining the multidimensional characteristic of system object, comprising:
Determine the primitive character value of the multidimensional characteristic of the system object;
The primitive character value is normalized, determines the normalization characteristic value of the multidimensional characteristic.
3. according to the method described in claim 2, it is characterized in that, determining the normalization characteristic value of the multidimensional characteristic, comprising:
In the importance correlation of the multidimensional characteristic and the system object, the normalization of the multidimensional characteristic is special
Value indicative is determined as the ratio of the primitive character value Yu the system object number;
In the importance negative correlation of the multidimensional characteristic and the system object, the normalization of the multidimensional characteristic is special
Value indicative is determined as 1 ratio for subtracting the primitive character value Yu the system object number.
4. according to the method in claim 2 or 3, which is characterized in that according to the multidimensional characteristic and feature of the system object
Weight determines the importance of the system object, comprising:
The system object is determined after being weighted according to normalization characteristic value of the feature weight to the multidimensional characteristic
Importance.
5. according to the method described in claim 4, it is characterized in that, according to the feature weight to the normalizing of the multidimensional characteristic
Change the importance that the system object is determined after characteristic value is weighted, comprising:
It is calculate by the following formula the importance values of the system object i:
Wherein, N indicates the characteristic dimension number of the system object i, wjIndicate the spy of j-th of feature in the multidimensional characteristic
Levy weight;f_scoreijIndicate the normalization characteristic value of j-th of feature of system object i.
6. -3,5 described in any item methods according to claim 1, which is characterized in that be based on training algorithm, determine the system
The corresponding feature weight of the multidimensional characteristic of object, comprising:
Obtain training sample;Wherein, the training sample include multiple system objects between any two importance first ratio
To result;
Using the training sample and minimal error rate the training algorithm feature weight corresponding to the multidimensional characteristic into
Row training, and then determine the corresponding feature weight of the multidimensional characteristic.
7. according to the method described in claim 6, it is characterized in that, being calculated using the training sample and minimal error rate training
The method feature weight corresponding to the multidimensional characteristic is trained, and then determines the corresponding feature power of the multidimensional characteristic
Weight, comprising:
Other features for the corresponding target signature weight of target signature in the multidimensional characteristic, in the multidimensional characteristic
In the case that corresponding feature weight is constant, determine multiple system objects under any value of target signature weight
Candidate importance;
According in the training sample first comparison result and the candidate importance determine the target signature power
Weight.
8. a kind of determining device of system object importance characterized by comprising
First determining module is configured to determine that the multidimensional characteristic of system object;
Second determining module is configured as determining the corresponding feature power of the multidimensional characteristic of the system object based on training algorithm
Weight;
Third determining module is configured as multidimensional characteristic and feature weight according to the system object, determines the system pair
The importance of elephant.
9. a kind of electronic equipment, which is characterized in that including memory and processor;Wherein,
The memory is for storing one or more computer instruction, wherein one or more computer instruction is by institute
Processor is stated to execute to realize the described in any item method and steps of claim 1-7.
10. a kind of computer readable storage medium, is stored thereon with computer instruction, which is characterized in that the computer instruction quilt
Claim 1-7 described in any item method and steps are realized when processor executes.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811565644.XA CN109658193A (en) | 2018-12-20 | 2018-12-20 | Determination method, apparatus, electronic equipment and the storage medium of system object importance |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811565644.XA CN109658193A (en) | 2018-12-20 | 2018-12-20 | Determination method, apparatus, electronic equipment and the storage medium of system object importance |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109658193A true CN109658193A (en) | 2019-04-19 |
Family
ID=66115402
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811565644.XA Pending CN109658193A (en) | 2018-12-20 | 2018-12-20 | Determination method, apparatus, electronic equipment and the storage medium of system object importance |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109658193A (en) |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103473317A (en) * | 2013-09-12 | 2013-12-25 | 百度在线网络技术(北京)有限公司 | Method and equipment for extracting keywords |
CN108038735A (en) * | 2017-12-26 | 2018-05-15 | 北京小度信息科技有限公司 | Data creation method and device |
US20180276291A1 (en) * | 2017-03-27 | 2018-09-27 | Alibaba Group Holding Limited | Method and device for constructing scoring model and evaluating user credit |
-
2018
- 2018-12-20 CN CN201811565644.XA patent/CN109658193A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103473317A (en) * | 2013-09-12 | 2013-12-25 | 百度在线网络技术(北京)有限公司 | Method and equipment for extracting keywords |
US20180276291A1 (en) * | 2017-03-27 | 2018-09-27 | Alibaba Group Holding Limited | Method and device for constructing scoring model and evaluating user credit |
CN108038735A (en) * | 2017-12-26 | 2018-05-15 | 北京小度信息科技有限公司 | Data creation method and device |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20190005518A1 (en) | Social network content item federation based on item utility value | |
US10572912B2 (en) | System and method for integrating retail price optimization for revenue and profit with business rules | |
US8374912B2 (en) | System and method for managing and optimizing advertising campaigns managed on the internet | |
Thürer et al. | Concerning workload control and order release: The pre‐shop pool sequencing decision | |
CN108492194A (en) | Products Show method, apparatus and storage medium | |
US20170323212A1 (en) | Agent aptitude prediction | |
US20080114778A1 (en) | System and method for generating a display of tags | |
US20200098031A1 (en) | Product recommending apparatus and non-transitory computer readable medium | |
JP2018077615A (en) | Advertising image generation device, advertising image generation method and program for advertising image generation device | |
CN112070545B (en) | Method, apparatus, medium, and electronic device for optimizing information reach | |
CN111340522A (en) | Resource recommendation method, device, server and storage medium | |
CN109934523A (en) | Information processing method and device | |
JP7206761B2 (en) | Information processing equipment | |
Hemmati et al. | A new decision making structure for managing arriving orders in MTO environments | |
US20050261881A1 (en) | Task-based design evaluation | |
CN116797235A (en) | Method and device for processing consumption information, storage medium and computer equipment | |
US10671601B2 (en) | Platform for consulting solution | |
CN109658193A (en) | Determination method, apparatus, electronic equipment and the storage medium of system object importance | |
Borissova et al. | Generalized approach to support business group decision-making by using of different strategies | |
CN101739621A (en) | Method and system for estimating technical asset value based on Web service | |
CN110163701A (en) | The method and apparatus of pushed information | |
US11544332B2 (en) | Bipartite graph construction | |
Osman et al. | Application of fuzzy logic for adaptive food recommendation | |
CN113344415A (en) | Deep neural network-based service distribution method, device, equipment and medium | |
US20080021795A1 (en) | Device for generating selection structures, for making selections according to selection structures, and for creating selection descriptions |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190419 |