CN110148023A - The electric power integral Method of Commodity Recommendation and system that logic-based returns - Google Patents
The electric power integral Method of Commodity Recommendation and system that logic-based returns Download PDFInfo
- Publication number
- CN110148023A CN110148023A CN201910403096.9A CN201910403096A CN110148023A CN 110148023 A CN110148023 A CN 110148023A CN 201910403096 A CN201910403096 A CN 201910403096A CN 110148023 A CN110148023 A CN 110148023A
- Authority
- CN
- China
- Prior art keywords
- user
- commodity
- accumulated point
- point exchanging
- integral
- 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 37
- 230000002596 correlated effect Effects 0.000 claims abstract description 44
- 238000001914 filtration Methods 0.000 claims abstract description 16
- 238000012549 training Methods 0.000 claims abstract description 15
- 238000007477 logistic regression Methods 0.000 claims abstract description 12
- 238000006243 chemical reaction Methods 0.000 claims abstract description 11
- 239000013065 commercial product Substances 0.000 claims description 28
- 230000010354 integration Effects 0.000 claims description 17
- 230000008859 change Effects 0.000 claims description 15
- PEDCQBHIVMGVHV-UHFFFAOYSA-N Glycerine Chemical compound OCC(O)CO PEDCQBHIVMGVHV-UHFFFAOYSA-N 0.000 claims description 8
- 238000011156 evaluation Methods 0.000 claims description 7
- 230000003542 behavioural effect Effects 0.000 claims description 6
- 238000013459 approach Methods 0.000 claims description 4
- 230000006399 behavior Effects 0.000 claims description 2
- 239000000284 extract Substances 0.000 claims description 2
- 239000000047 product Substances 0.000 claims description 2
- 238000009825 accumulation Methods 0.000 claims 1
- 230000006870 function Effects 0.000 description 9
- 230000000694 effects Effects 0.000 description 6
- 238000002474 experimental method Methods 0.000 description 5
- 238000012360 testing method Methods 0.000 description 5
- 230000008901 benefit Effects 0.000 description 4
- 230000000875 corresponding effect Effects 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000000052 comparative effect Effects 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 235000010894 Artemisia argyi Nutrition 0.000 description 1
- 244000030166 artemisia Species 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000002360 explosive Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 239000004984 smart glass Substances 0.000 description 1
- 230000002123 temporal effect Effects 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
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
-
- 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
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0207—Discounts or incentives, e.g. coupons or rebates
- G06Q30/0222—During e-commerce, i.e. online transactions
-
- 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
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0207—Discounts or incentives, e.g. coupons or rebates
- G06Q30/0224—Discounts or incentives, e.g. coupons or rebates based on user history
-
- 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
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0207—Discounts or incentives, e.g. coupons or rebates
- G06Q30/0226—Incentive systems for frequent usage, e.g. frequent flyer miles programs or point systems
-
- 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]
- G06Q30/0631—Item recommendations
-
- 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
- G06Q50/06—Energy or water supply
Landscapes
- Business, Economics & Management (AREA)
- Finance (AREA)
- Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- Strategic Management (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Theoretical Computer Science (AREA)
- Marketing (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Primary Health Care (AREA)
- Tourism & Hospitality (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The electric power integral Method of Commodity Recommendation and system returned the present disclosure discloses logic-based, include: to obtain to have integral and the most correlated characteristic data of the positive sample user that once crosses commodity with accumulated point exchanging for every a kind of user, also obtain have integral and unused accumulated point exchanging cross commodity negative sample user most correlated characteristic data;Using logistic regression algorithm, label establishes potential customers' prediction model as training set data whether using the most correlated characteristic data, the most correlated characteristic data of negative sample user and accumulated point exchanging of positive sample user;Based on potential customers' prediction model, according to the historical integral time record of conversion of user to be predicted, predict that user carries out the probability of accumulated point exchanging;It is greater than the user of given threshold for probability, is considered as potential accumulated point exchanging user, uses collaborative filtering for potential accumulated point exchanging user Recommendations.
Description
Technical field
This disclosure relates to electric power integral Method of Commodity Recommendation that personalized recommendation field more particularly to logic-based return and
System.
Background technique
The statement of this part is only to refer to background technique relevant to the disclosure, not necessarily constitutes the prior art.
In implementing the present disclosure, following technical problem exists in the prior art in inventor:
In recent years, the high speed development of information technology causes the quantity of Internet user to be continuously increased, data volume exponentially type
Explosive growth, people have come into the epoch of an information overload, how to have obtained in the information of magnanimity under big data background
The information that user wants has become an important research topic, and proposed algorithm overcomes the important skill of information overload as one
Art has been widely used for e-commerce field, by academia and industry long-term research and application, comparative maturity,
Its economic benefit generated is huge.
What is be most widely used in recommender system at present surely belongs to collaborative filtering, including based on neighbours and model
Two class methods, the method core based on neighbours are the similarities calculated between user or between article to carry out pushing away for next step
Recommend work.Method core content based on model is to convert different models, example for user-article relationship assessment data
Such as factorization, Bayesian network model are recommended by these models to user accordingly.User interest has the time
Property, the recommender system of current main-stream usually lays particular emphasis on consideration user interest preference when recommending to user, emerging for user
The considerations of interest variation, then shows slightly insufficient.Different users has different behavioural habits, is also in this way, passing in e-commerce field
The recommender system of system does not account for the behavioural habits of user, is all calculated and is recommended, this is virtually increasing system
Burden, but also recommend Objective it is not strong.
Summary of the invention
In order to solve the deficiencies in the prior art, present disclose provides the electric power that logic-based returns to integrate Method of Commodity Recommendation
And system filters out the user that can be carried out accumulated point exchanging by a kind of potential customers' prediction model, to make recommended work more
With Objective, the burden of recommender system also can reduce, recommend to use collaborative filtering, in traditional collaborative filtering
On the basis of binding time factor be target user recommend.
In a first aspect, present disclose provides the electric power that logic-based returns to integrate Method of Commodity Recommendation;
The electric power that logic-based returns integrates Method of Commodity Recommendation, comprising:
User is clustered, different user groups is obtained;
The most correlated characteristic for having the positive sample user for integrating and once crossing commodity with accumulated point exchanging is obtained for every a kind of user
Data, also obtain have integral and unused accumulated point exchanging cross commodity negative sample user most correlated characteristic data;
Using logistic regression algorithm, with the most correlated characteristic data of positive sample user, the most correlated characteristic of negative sample user
Label establishes potential customers' prediction model as training set data whether data and accumulated point exchanging;
Based on potential customers' prediction model, according to the historical integral time record of conversion of user to be predicted, predict that user accumulates
Divide the probability exchanged;
For probability be greater than given threshold user, be considered as potential accumulated point exchanging user, use collaborative filtering for
Potential accumulated point exchanging user's Recommendations.
Second aspect, present disclose provides the electric power that logic-based returns to integrate commercial product recommending system;
The electric power that logic-based returns integrates commercial product recommending system, comprising:
User's categorization module, clusters user, obtains different user groups;
Most correlated characteristic obtains module, and being configured as obtaining for every a kind of user has integral and once with accumulated point exchanging mistake
The most correlated characteristic data of the positive sample user of commodity, also acquisition have integral and unused accumulated point exchanging crosses the negative sample user of commodity
Most correlated characteristic data;
Potential customers' prediction model constructs module, is configured as using logistic regression algorithm, most with positive sample user
Label is built as training set data whether correlated characteristic data, the most correlated characteristic data of negative sample user and accumulated point exchanging
Vertical potential customers' prediction model;
Accumulated point exchanging user in predicting module is configured as based on potential customers' prediction model, according to user's to be predicted
Historical integral time record of conversion, prediction user carry out the probability of accumulated point exchanging;
Commercial product recommending module is configured as being greater than probability the user of given threshold, is considered as potential accumulated point exchanging
User uses collaborative filtering for potential accumulated point exchanging user Recommendations.
The third aspect, the disclosure additionally provide a kind of electronic equipment, including memory and processor and are stored in storage
The computer instruction run on device and on a processor when the computer instruction is run by processor, completes first aspect institute
The step of stating method.
Fourth aspect, the disclosure additionally provide a kind of computer readable storage medium, described for storing computer instruction
When computer instruction is executed by processor, complete first aspect the method the step of.
Beneficial effects of the present invention:
1. potential customers' prediction model that logic-based regression algorithm is established, solves potential accumulated point exchanging to a certain extent
The determination problem of client, makes full use of sample information, keeps the result excavated also reliable, poly- by carrying out layering to user
Class can not only refine the classification of user, while can also obtain more accurate user most correlated characteristic information, logistic regression
The use of algorithm can quickly determine that user carries out a possibility that exchanging integral, to make next recommended work with more needle
To property, potential customers' prediction model emphasis is user's cluster, and core is logistic regression prediction, can when extracting user's characteristic information
Not to be restricted to one pattern, user data is enriched according to the actual situation to obtain different user most related data.
2. recommending aspect using collaborative filtering, the commodity operated according to target user pass through calculating
Time-based user is to the preference of commodity so that it is determined that the potential preference article of target user, is pushed away according to article similarity
It recommends list and returns to user.
3, entire model includes two parts: potential customers' prediction and collaborative filtering recommending;Potential customers' prediction model advantage
It is, with the increase of data volume, not will lead to calculation amount compared to other models and uncomplicated and exponentially rise, this point is full
Requirement of the pedal system to the time;The advantage of part is recommended to be not needing to compute repeatedly when calculating commodity similarity, it is only necessary to
It just will do it calculating in system article change, equally meet system real time requirement.
Detailed description of the invention
The accompanying drawings constituting a part of this application is used to provide further understanding of the present application, and the application's shows
Meaning property embodiment and its explanation are not constituted an undue limitation on the present application for explaining the application.
Fig. 1 potential customers' prediction model figure;
Fig. 2 database diagram;
Data sample figure when Fig. 3 is tested;
Fig. 4 recommender system illustraton of model;
Fig. 5 user's redemption data ratio chart;
Whether there is or not potential customers' prediction effect comparisons that correlated characteristic is analyzed by Fig. 6;
Fig. 7 recommends accuracy rate;
Fig. 8 recommends recall rate.
Specific embodiment
It is noted that described further below be all exemplary, it is intended to provide further instruction to the application.Unless another
It indicates, all technical and scientific terms used herein has usual with the application person of an ordinary skill in the technical field
The identical meanings of understanding.
It should be noted that term used herein above is merely to describe specific embodiment, and be not intended to restricted root
According to the illustrative embodiments of the application.As used herein, unless the context clearly indicates otherwise, otherwise singular
Also it is intended to include plural form, additionally, it should be understood that, when in the present specification using term "comprising" and/or " packet
Include " when, indicate existing characteristics, step, operation, device, component and/or their combination.
Embodiment one present embodiments provides the electric power integral Method of Commodity Recommendation of logic-based recurrence;
The electric power that logic-based returns integrates Method of Commodity Recommendation, comprising:
S0: clustering user, obtains different user groups;
S1: the most related spy for having the positive sample user for integrating and once crossing commodity with accumulated point exchanging is obtained for every a kind of user
Levy data, also obtain have integral and unused accumulated point exchanging cross commodity negative sample user most correlated characteristic data;
S2: logistic regression algorithm is used, with the most correlated characteristic data of positive sample user, the most related spy of negative sample user
Label establishes potential customers' prediction model as training set data whether levying data and accumulated point exchanging;
S3: being based on potential customers' prediction model, according to the historical integral time record of conversion of user to be predicted, predicts that user carries out
The probability of accumulated point exchanging;
S4: it is greater than the user of given threshold for probability, is considered as potential accumulated point exchanging user, using collaborative filtering
For potential accumulated point exchanging user Recommendations.
It is described that user is clustered as one or more embodiments, different user groups is obtained, is referred to:
Based on the evaluation index of accumulated point exchanging user, classified to user using cluster mode;
To the historical integral time record of conversion of every a kind of user, carried out using feature selection approach based on mutual information most related
Feature selecting.
As one or more embodiments, the evaluation index based on accumulated point exchanging user uses cluster side to user
Formula is classified;Specific steps include:
The evaluation index of accumulated point exchanging user, comprising: the year integration change volume and moon integration change volume of accumulated point exchanging user;
The year integration change volume and moon integration change volume of acquired integrated exchange user;
User is exchanged to year integration using FCM Algorithms to cluster, and obtains r cluster;
User is exchanged to moon integration using FCM Algorithms to cluster, and obtains s cluster;
User is divided into r × s class.
As one or more embodiments, the historical integral time record of conversion, comprising: user integral cumulative balance, user
Accumulated point exchanging information, user integral exchange the time, user integral exchanges number, user integral exchanges frequency and last time integrates
Time interval of the exchange behavior time of origin to current time.
As one or more embodiments, the selection of most correlated characteristic is carried out using feature selection approach based on mutual information,
It is to choose the maximum preceding k feature of mutual information as such user most correlated characteristic.
As one or more embodiments, use collaborative filtering for potential accumulated point exchanging user Recommendations, specifically
Step includes:
The history goods browse behavioral data of potential user is acquired, extracts historical viewings item property, history buys commodity
The price of attribute and history purchase commodity;
Calculate the first similarity that store has the attribute of commodity and user's history browsing commodity;First similarity is greater than
The store of given threshold has commodity and stores into the first commercial product recommending set;
Calculate the second similarity of the attribute of the commodity and user's history purchase commodity in the first commercial product recommending set;By
Two similarities are greater than the commodity in the first commercial product recommending set of given threshold and store into the second commercial product recommending set;
Calculate the difference of the price of the commodity and user's history purchase commodity in the second commercial product recommending set;Difference is less than
Commodity in second commercial product recommending set of given threshold are stored into third commercial product recommending set;
Calculate the time difference at time point and current point in time that the commodity in third commercial product recommending set are browsed by user, base
Preference of the user to each commodity in third commercial product recommending set, time difference smaller expression preference are calculated in the time difference
It is bigger;
By the commodity in third commercial product recommending set, exported according to preference descending.
It can equally be carried out according to the actual situation when calculating commodity similarity substantial, can such as increase information such as user
Monthly average integral increment, the every monthly average of user exchange number, exchange time interval etc., to obtain more accurate digging
Dig effect.The scalability of model is also a big advantage of this model more by force, has better adaptability.
Similarity between article is calculated using Euclidean distance, independent variable is the price and classification of commodity, from calculating
On say, not needing complicated calculating, improving algorithm operational efficiency, precision aspect meets subsequent recommendation work, in recommendation results
Influence of the time factor to user interest is considered, the content recommended is made to have more reasonability.
As one or more embodiments, history goods browse behavioral data, comprising: basic information, the Yong Husuo of user
The attribute for the commodity that the attribute for the commodity that the attribute of the commodity browsed, user seeked advice from, user collected, user's browsing
The commodity price that the number of commodity and user bought.
It is to be understood that the multidimensional evaluation index form that user exchanges integral is as follows:
Wherein α and β is temporal aspect vector, respectively indicates the year integration change volume time series data and moon integration of user
Change volume time series data, i indicate user, VDIndicate user data.
The selection calculation formula of most correlated characteristic is as follows:
Inf in above formulaiIndicate each feature, yiIndicate whether integral is exchanged.P (X=infi, Y=yi) indicate infiAnd yi
While appear in probability in entire data set, each feature is ranked up by the calculating of above formula, and choose mutual information
Most correlated characteristic information of the maximum preceding k feature as such user logic regression modeling.
For different user groups, chosen on the basis of potential customers predict most correlated characteristic information data corresponding
Data form the data set of experiment, and treated information is divided into training set and test set according to the ratio of 4:1, for pair
The training and test of model.
Model training.Using logistic regression method, the probability that user carries out accumulated point exchanging is set as P, then
Wherein, f=UT*X+WT*Y+AT* the condition that T, U, W, A meet is U=(u1,u2,u3), W=(w1), A=(a1,a2,
a3), U, W, A are parameter vector, and wherein number of parameters can change according to the difference of group's most correlated characteristic, initial value by
Artificial setting, experimental result are adjusted.The likelihood letter function that sample occurs is constructed, and is counted using gradient rising
Calculate the value for obtaining parameters.
The similarity between commodity is calculated, specific formula for calculation is as follows:
Above formula indicates the similarity of commodity A and commodity B, βiIt is weight parameter,Can according to the actual situation with
And training is adjusted;aiAnd biValue description be shown in Table 1.
1 commodity part information table of table
After obtaining the similarity between commodity, calculate time-based user to item according to the time parameter of table 2
Purpose preference.
2 user preference information of table
User interest changing rule is that the interest of user in a short time is constant, and the interest of long-term interior user is in non-linear forgetting,
Therefore the universal forgetting law according to Chinese mugwort guest great this forgetting curve and people is selected power function as the function of time, is ground according to correlation
The function for studying carefully use is as follows:
F (u, i)=0.318 × (T0-Tui)-0.125
T in above formula0Indicate the time point currently recommended, TuiIndicate the time point of user u operation i, (T0-TuiBoth) indicate
Between number of days, f (u, i) indicates user u to the preference of i.
It is these three types of by calculating a user's exchange of acquisition K (value of K determines according to the actual situation), collection and consulting
In commodity the time-based highest commodity of user preference degree as user potential preference commodity and be organized into set m ', point
Not Huo Qu with each the most similar commodity of commodity in set m ' and be integrated into recommendation list, the number of commodity in recommendation list
It can determine according to the actual situation.
The embodiment of the present invention provides a kind of electric power integral store proposed algorithm that logic-based returns.Traditional proposed algorithm
Objective is not strong, and when integral store number of users is huge, will lead to recommender system, over-burden, expends a large amount of financial resources object
Power can determine a possibility that target user carries out accumulated point exchanging by using logistic regression algorithm to a certain extent, thus
Targetedly recommend.Recommendation for end article then uses the collaborative filtering of current comparative maturity, on this basis
Reference time information is recommended accordingly, to better conform to the variation of user interest, improves the effect of recommendation.
It is the illustraton of model that the present invention is recommended on the basis of logistic regression algorithm with reference to Fig. 4, Fig. 4, step includes such as
Under:
A. based on Shandong electric power integral store data, database diagram is required to search according to Fig.2,
User information data, be broadly divided into four parts: basic information, integration information, redemption information and the time letter of store member
Breath, and summarize arrangement and obtain information summary sheet and merchandise news table (content is commodity ID, merchandise classification, commodity price, click
Number).
B. according to the year integration change volume time series data of user and moon integration change volume time series data to user
Fuzzy C-means clustering by different level is carried out, clusters number c=3 is initialized, fog-level Coefficient m=2 obtain 9 user groups,
The most correlated characteristic data for obtaining different user group are analyzed by most correlated characteristic, these data are further processed simultaneously
Screening obtains experiment data set.
C. the nearly 2 years records in family are taken as data set, data set is divided into five portions according to the difference of user group
Point, separately include five training sets and test set, wherein user's redemption data proportion in five parts of training sets and test set
As shown in Figure 5.
D. it is trained and is obtained corresponding respectively using potential customers' prediction model of the training set to different user group
Parameter.Target user can be predicted after obtaining relevant parameter.Calculated by Sigmoid function target user into
A possibility that row is exchanged, value determine that it can be carried out exchange when being greater than 0.5, it is determined when value is less than 0.5 not will do it exchange, real
Different groups user most correlated characteristic information is different in data set used when testing, by taking one type as an example, most related spy
Reference breath includes: store member age, binding Electricity customers number, integral payment times, last time exchange to the now time
Interval, collection and consulting commodity number, accumulated point exchanging frequency, balance of points, summarize as shown in table 3, data sample is as shown in Figure 3.
Certain user group of table 3 most correlated characteristic information summary sheet
Wherein four dimensions, six dimensions, the data of eight dimensions are respectively adopted and carry out the training of potential customers' prediction model,
The experimental result of each group of comprehensive analysis shows as follows: when dimension value is 4, experiment overall accuracy in 39%~
45%, when dimension value is 6, experiment overall accuracy is in 40%~60%, when experiment dimension value is 8, overall accuracy
It in 55%~78%, thus sees, the information content for increasing potential customers' prediction model can promote recommendation to a certain extent
Effect.Under different user group, whether there is or not potential customers' prediction effects of correlated characteristic analysis to compare as shown in fig. 6, can see
Out, the latent objective digging office model established by correlated characteristic analysis can preferably excavate potential accumulated point exchanging user.
E. the present invention uses the collaborative filtering based on article.It calculates article similarity and is integrated into article similarity
Table, for the store member of existing record of conversion, seek user preference three classes commodity, be respectively collecting commodities, exchange commodity
With consulting commodity and corresponding operation time, time-based user is obtained to the preference of commodity, K before choosing by calculating
Commodity obtain this most similar commodity of K commodity by article similarity table respectively and are aggregated into as user preference commodity
Recommendation list (value of K depends on the circumstances), the general effect of recommendation is as shown in Figure 7 and Figure 8.
Embodiment two: the electric power integral commercial product recommending system of logic-based recurrence is present embodiments provided;
The electric power that logic-based returns integrates commercial product recommending system, comprising:
User's categorization module, clusters user, obtains different user groups;
Most correlated characteristic obtains module, and being configured as obtaining for every a kind of user has integral and once with accumulated point exchanging mistake
The most correlated characteristic data of the positive sample user of commodity, also obtaining has integral and uses from the negative sample that unused accumulated point exchanging crosses commodity
The most correlated characteristic data at family;
Potential customers' prediction model constructs module, is configured as using logistic regression algorithm, most with positive sample user
Label is built as training set data whether correlated characteristic data, the most correlated characteristic data of negative sample user and accumulated point exchanging
Vertical potential customers' prediction model, as shown in Figure 1;
Accumulated point exchanging user in predicting module is configured as based on potential customers' prediction model, according to user's to be predicted
Historical integral time record of conversion, prediction user carry out the probability of accumulated point exchanging;
Commercial product recommending module is configured as being greater than probability the user of given threshold, is considered as potential accumulated point exchanging
User uses collaborative filtering for potential accumulated point exchanging user Recommendations.
Embodiment three: the present embodiment additionally provides a kind of electronic equipment, including memory and processor and is stored in
The computer instruction run on reservoir and on a processor, when the computer instruction is run by processor, in Method Of Accomplishment
Each operation, for sake of simplicity, details are not described herein.
The electronic equipment can be mobile terminal and immobile terminal, and immobile terminal includes desktop computer, move
Dynamic terminal includes smart phone (Smart Phone, such as Android phone, IOS mobile phone), smart glasses, smart watches, intelligence
The mobile internet device that energy bracelet, tablet computer, laptop, personal digital assistant etc. can carry out wireless communication.
It should be understood that in the disclosure, which can be central processing unit CPU, which, which can be said to be, can be it
His general processor, digital signal processor DSP, application-specific integrated circuit ASIC, ready-made programmable gate array FPGA or other
Programmable logic device, discrete gate or transistor logic, discrete hardware components etc..General processor can be micro process
Device or the processor are also possible to any conventional processor etc..
The memory may include read-only memory and random access memory, and to processor provide instruction and data,
The a part of of memory can also include non-volatile RAM.For example, memory can be with the letter of storage device type
Breath.
During realization, each step of the above method can by the integrated logic circuit of the hardware in processor or
The instruction of software form is completed.The step of method in conjunction with disclosed in the disclosure, can be embodied directly in hardware processor and execute
At, or in processor hardware and software module combination execute completion.Software module can be located at random access memory, dodge
It deposits, this fields are mature deposits for read-only memory, programmable read only memory or electrically erasable programmable memory, register etc.
In storage media.The storage medium is located at memory, and processor reads the information in memory, completes the above method in conjunction with its hardware
The step of.To avoid repeating, it is not detailed herein.Those of ordinary skill in the art may be aware that in conjunction with institute herein
Each exemplary unit, that is, algorithm steps of disclosed embodiment description, can be hard with electronic hardware or computer software and electronics
The combination of part is realized.These functions are implemented in hardware or software actually, the specific application depending on technical solution
And design constraint.Professional technician can realize described function using distinct methods to each specific application
Can, but this realization is it is not considered that exceed scope of the present application.
It is apparent to those skilled in the art that for convenience and simplicity of description, the system of foregoing description,
The specific work process of device and unit, can refer to corresponding processes in the foregoing method embodiment, and details are not described herein.
In several embodiments provided herein, it should be understood that disclosed systems, devices and methods, it can be with
It realizes in other way.For example, the apparatus embodiments described above are merely exemplary, for example, the unit
It divides, only a kind of division of logic function, there may be another division manner in actual implementation, such as multiple units or group
Part can be combined or can be integrated into another system, or some features can be ignored or not executed.In addition, showing
The mutual coupling or direct-coupling or communication connection shown or discussed can be through some interfaces, device or unit
Indirect coupling or communication connection, can be electrically, mechanical or other forms.
It, can be with if the function is realized in the form of SFU software functional unit and when sold or used as an independent product
It is stored in a computer readable storage medium.Based on this understanding, the technical solution of the application is substantially right in other words
The part of part or the technical solution that the prior art contributes can be embodied in the form of software products, the calculating
Machine software product is stored in a storage medium, including some instructions are used so that a computer equipment (can be individual
Computer, server or network equipment etc.) execute each embodiment the method for the application all or part of the steps.And it is preceding
The storage medium stated includes: USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory
The various media that can store program code such as (RAM, Random Access Memory), magnetic or disk.
The foregoing is merely preferred embodiment of the present application, are not intended to limit this application, for the skill of this field
For art personnel, various changes and changes are possible in this application.Within the spirit and principles of this application, made any to repair
Change, equivalent replacement, improvement etc., should be included within the scope of protection of this application.
Claims (10)
1. the electric power that logic-based returns integrates Method of Commodity Recommendation, characterized in that include:
User is clustered, different user groups is obtained;
The most correlated characteristic data for having the positive sample user for integrating and once crossing commodity with accumulated point exchanging are obtained for every a kind of user,
Also obtain have integral and unused accumulated point exchanging cross commodity negative sample user most correlated characteristic data;
Using logistic regression algorithm, with the most correlated characteristic data of positive sample user, the most correlated characteristic data of negative sample user,
And label establishes potential customers' prediction model as training set data whether accumulated point exchanging;
Based on potential customers' prediction model, according to the historical integral time record of conversion of user to be predicted, predict that user carries out integral and converts
The probability changed;
It is greater than the user of given threshold for probability, is considered as potential accumulated point exchanging user, uses collaborative filtering to be potential
Accumulated point exchanging user's Recommendations.
2. the method as described in claim 1, characterized in that it is described that user is clustered, different user groups is obtained, is
Refer to:
Based on the evaluation index of accumulated point exchanging user, classified to user using cluster mode;
To the historical integral time record of conversion of every a kind of user, most correlated characteristic is carried out using feature selection approach based on mutual information
Selection.
3. method according to claim 2, characterized in that the evaluation index based on accumulated point exchanging user adopts user
Classified with cluster mode;Specific steps include:
The evaluation index of accumulated point exchanging user, comprising: the year integration change volume and moon integration change volume of accumulated point exchanging user;
The year integration change volume and moon integration change volume of acquired integrated exchange user;
User is exchanged to year integration using FCM Algorithms to cluster, and obtains r cluster;
User is exchanged to moon integration using FCM Algorithms to cluster, and obtains s cluster;
User is divided into r × s class.
4. the method as described in claim 1, characterized in that the historical integral time record of conversion, comprising: more than user integral accumulation
Volume, user integral redemption information, user integral exchange the time, user integral exchanges number, user integral exchanges frequency or last
Time interval of the accumulated point exchanging behavior time of origin to current time.
5. the method as described in claim 1, characterized in that carry out most related spy using feature selection approach based on mutual information
Sign selection is to choose the maximum preceding k feature of mutual information as such user most correlated characteristic.
6. the method as described in claim 1, characterized in that collaborative filtering is used to recommend quotient for potential accumulated point exchanging user
Product, specific steps include:
The history goods browse behavioral data of potential user is acquired, extracts historical viewings item property, history buys item property
With the price of history purchase commodity;
Calculate the first similarity that store has the attribute of commodity and user's history browsing commodity;First similarity is greater than setting
The store of threshold value has commodity and stores into the first commercial product recommending set;
Calculate the second similarity of the attribute of the commodity and user's history purchase commodity in the first commercial product recommending set;By the second phase
The commodity being greater than in the first commercial product recommending set of given threshold like degree are stored into the second commercial product recommending set;
Calculate the difference of the price of the commodity and user's history purchase commodity in the second commercial product recommending set;Difference is less than setting
Commodity in second commercial product recommending set of threshold value are stored into third commercial product recommending set;
The time difference at time point and current point in time that the commodity in third commercial product recommending set are browsed by user is calculated, when being based on
Between difference calculate user to the preference of each commodity in third commercial product recommending set, the time difference, smaller expression preference was got over
Greatly;
By the commodity in third commercial product recommending set, exported according to preference descending.
7. method as claimed in claim 6, characterized in that history goods browse behavioral data, comprising: the basis letter of user
The category for the commodity that the attribute for the commodity that the attribute for the commodity that breath, user browsed, user seeked advice from, user collected
Property, user browse the number and the commodity price bought of user of commodity.
8. the electric power that logic-based returns integrates commercial product recommending system, characterized in that include:
User's categorization module, clusters user, obtains different user groups;
Most correlated characteristic obtains module, and being configured as obtaining for every a kind of user has integral and once cross commodity with accumulated point exchanging
Positive sample user most correlated characteristic data, also obtain have integral and unused accumulated point exchanging cross the negative sample user of commodity most
Correlated characteristic data;
Potential customers' prediction model constructs module, is configured as using logistic regression algorithm, with the most related of positive sample user
Label is established latent as training set data whether characteristic, the most correlated characteristic data of negative sample user and accumulated point exchanging
In customer predictability model;
Accumulated point exchanging user in predicting module is configured as based on potential customers' prediction model, according to the history of user to be predicted
Accumulated point exchanging record, prediction user carry out the probability of accumulated point exchanging;
Commercial product recommending module is configured as being greater than probability the user of given threshold, is considered as potential accumulated point exchanging user,
Use collaborative filtering for potential accumulated point exchanging user Recommendations.
9. a kind of electronic equipment, characterized in that on a memory and on a processor including memory and processor and storage
The computer instruction of operation when the computer instruction is run by processor, is completed described in any one of claim 1-7 method
Step.
10. a kind of computer readable storage medium, characterized in that for storing computer instruction, the computer instruction is located
When managing device execution, step described in any one of claim 1-7 method is completed.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910403096.9A CN110148023A (en) | 2019-05-15 | 2019-05-15 | The electric power integral Method of Commodity Recommendation and system that logic-based returns |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910403096.9A CN110148023A (en) | 2019-05-15 | 2019-05-15 | The electric power integral Method of Commodity Recommendation and system that logic-based returns |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110148023A true CN110148023A (en) | 2019-08-20 |
Family
ID=67595457
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910403096.9A Pending CN110148023A (en) | 2019-05-15 | 2019-05-15 | The electric power integral Method of Commodity Recommendation and system that logic-based returns |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110148023A (en) |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111931065A (en) * | 2020-09-03 | 2020-11-13 | 猪八戒股份有限公司 | Business opportunity recommendation method, system, electronic device and medium based on LSTM model |
CN112036951A (en) * | 2020-09-03 | 2020-12-04 | 猪八戒股份有限公司 | Business opportunity recommendation method, system, electronic device and medium based on CNN model |
CN112948687A (en) * | 2021-03-25 | 2021-06-11 | 重庆高开清芯智联网络科技有限公司 | Node message recommendation method based on name card file characteristics |
CN113222632A (en) * | 2020-02-04 | 2021-08-06 | 北京京东振世信息技术有限公司 | Object mining method and device |
CN113781134A (en) * | 2020-07-28 | 2021-12-10 | 北京沃东天骏信息技术有限公司 | Item recommendation method and device and computer-readable storage medium |
CN114223012A (en) * | 2019-10-31 | 2022-03-22 | 深圳市欢太科技有限公司 | Push object determination method and device, terminal equipment and storage medium |
CN114612159A (en) * | 2022-03-23 | 2022-06-10 | 何英秀 | Online mall platform intelligent management method, system and computer storage medium |
CN115481813A (en) * | 2022-10-09 | 2022-12-16 | 武汉理工大学 | New energy bus energy consumption prediction method based on machine learning |
CN116089401A (en) * | 2023-02-17 | 2023-05-09 | 国网浙江省电力有限公司营销服务中心 | User data management method and system |
CN116228342A (en) * | 2022-10-24 | 2023-06-06 | 广州麦乐数字科技有限公司 | Commodity recommendation method and device and computer readable storage medium |
CN116957657A (en) * | 2023-06-26 | 2023-10-27 | 北京信大融金教育科技有限公司 | Bank point processing method, device, equipment and medium based on operation model |
CN117114760A (en) * | 2023-05-04 | 2023-11-24 | 浙江卡赢信息科技有限公司 | Intelligent recommendation system and method for point exchange based on user consumption behavior analysis |
CN117132368A (en) * | 2023-10-26 | 2023-11-28 | 深圳企域数字科技有限公司 | Novel media intelligent marketing platform based on AI |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105512768A (en) * | 2015-12-14 | 2016-04-20 | 上海交通大学 | User electricity consumption relevant factor identification and electricity consumption quantity prediction method under environment of big data |
CN107133843A (en) * | 2017-04-25 | 2017-09-05 | 丹露成都网络技术有限公司 | A kind of Method of Commodity Recommendation based on collaborative filtering |
US20180052906A1 (en) * | 2016-08-22 | 2018-02-22 | Facebook, Inc. | Systems and methods for recommending content items |
CN108334887A (en) * | 2017-01-19 | 2018-07-27 | 腾讯科技(深圳)有限公司 | A kind of user's choosing method and device |
-
2019
- 2019-05-15 CN CN201910403096.9A patent/CN110148023A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105512768A (en) * | 2015-12-14 | 2016-04-20 | 上海交通大学 | User electricity consumption relevant factor identification and electricity consumption quantity prediction method under environment of big data |
US20180052906A1 (en) * | 2016-08-22 | 2018-02-22 | Facebook, Inc. | Systems and methods for recommending content items |
CN108334887A (en) * | 2017-01-19 | 2018-07-27 | 腾讯科技(深圳)有限公司 | A kind of user's choosing method and device |
CN107133843A (en) * | 2017-04-25 | 2017-09-05 | 丹露成都网络技术有限公司 | A kind of Method of Commodity Recommendation based on collaborative filtering |
Cited By (19)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114223012A (en) * | 2019-10-31 | 2022-03-22 | 深圳市欢太科技有限公司 | Push object determination method and device, terminal equipment and storage medium |
CN113222632A (en) * | 2020-02-04 | 2021-08-06 | 北京京东振世信息技术有限公司 | Object mining method and device |
CN113781134A (en) * | 2020-07-28 | 2021-12-10 | 北京沃东天骏信息技术有限公司 | Item recommendation method and device and computer-readable storage medium |
CN111931065A (en) * | 2020-09-03 | 2020-11-13 | 猪八戒股份有限公司 | Business opportunity recommendation method, system, electronic device and medium based on LSTM model |
CN112036951A (en) * | 2020-09-03 | 2020-12-04 | 猪八戒股份有限公司 | Business opportunity recommendation method, system, electronic device and medium based on CNN model |
CN112948687A (en) * | 2021-03-25 | 2021-06-11 | 重庆高开清芯智联网络科技有限公司 | Node message recommendation method based on name card file characteristics |
CN114612159A (en) * | 2022-03-23 | 2022-06-10 | 何英秀 | Online mall platform intelligent management method, system and computer storage medium |
CN114612159B (en) * | 2022-03-23 | 2022-11-22 | 鼎趣(上海)科技有限公司 | Online mall platform intelligent management method, system and computer storage medium |
CN115481813A (en) * | 2022-10-09 | 2022-12-16 | 武汉理工大学 | New energy bus energy consumption prediction method based on machine learning |
CN115481813B (en) * | 2022-10-09 | 2024-05-31 | 武汉理工大学 | New energy bus energy consumption prediction method based on machine learning |
CN116228342A (en) * | 2022-10-24 | 2023-06-06 | 广州麦乐数字科技有限公司 | Commodity recommendation method and device and computer readable storage medium |
CN116228342B (en) * | 2022-10-24 | 2024-02-06 | 广州麦乐数字科技有限公司 | Commodity recommendation method and device and computer readable storage medium |
CN116089401A (en) * | 2023-02-17 | 2023-05-09 | 国网浙江省电力有限公司营销服务中心 | User data management method and system |
CN116089401B (en) * | 2023-02-17 | 2023-09-05 | 国网浙江省电力有限公司营销服务中心 | User data management method and system |
CN117114760A (en) * | 2023-05-04 | 2023-11-24 | 浙江卡赢信息科技有限公司 | Intelligent recommendation system and method for point exchange based on user consumption behavior analysis |
CN117114760B (en) * | 2023-05-04 | 2024-05-28 | 浙江卡赢信息科技有限公司 | Intelligent recommendation system and method for point exchange based on user consumption behavior analysis |
CN116957657A (en) * | 2023-06-26 | 2023-10-27 | 北京信大融金教育科技有限公司 | Bank point processing method, device, equipment and medium based on operation model |
CN117132368A (en) * | 2023-10-26 | 2023-11-28 | 深圳企域数字科技有限公司 | Novel media intelligent marketing platform based on AI |
CN117132368B (en) * | 2023-10-26 | 2024-01-30 | 深圳企域数字科技有限公司 | Novel media intelligent marketing platform based on AI |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110148023A (en) | The electric power integral Method of Commodity Recommendation and system that logic-based returns | |
CN110503531B (en) | Dynamic social scene recommendation method based on time sequence perception | |
CN105224699B (en) | News recommendation method and device | |
CN108090800B (en) | Game prop pushing method and device based on player consumption potential | |
Shinde et al. | Hybrid personalized recommender system using centering-bunching based clustering algorithm | |
CN110427560A (en) | A kind of model training method and relevant apparatus applied to recommender system | |
CN109948121A (en) | Article similarity method for digging, system, equipment and storage medium | |
CN110457577B (en) | Data processing method, device, equipment and computer storage medium | |
CN110489642A (en) | Method of Commodity Recommendation, system, equipment and the medium of Behavior-based control signature analysis | |
CN109036577A (en) | Diabetic complication analysis method and device | |
CN116431931B (en) | Real-time incremental data statistical analysis method | |
Nazemi et al. | Visual analytics for analyzing technological trends from text | |
CN108615177A (en) | Electric terminal personalized recommendation method based on weighting extraction interest-degree | |
CN107908616A (en) | The method and apparatus of anticipation trend word | |
CN106294882A (en) | Data digging method and device | |
CN116205675A (en) | Data acquisition method and device based on thread division | |
CN111861605A (en) | Business object recommendation method | |
CN110852785A (en) | User grading method, device and computer readable storage medium | |
CN113239182A (en) | Article recommendation method and device, computer equipment and storage medium | |
Xi et al. | Multi-level interaction reranking with user behavior history | |
Vaganov et al. | Forecasting purchase categories with transition graphs using financial and social data | |
Wu et al. | RETRACTED ARTICLE: Artificial neural network based high dimensional data visualization technique for interactive data exploration in E-commerce | |
Sagar et al. | Analysis of prediction techniques based on classification and regression | |
CN110287373A (en) | Collaborative filtering film recommended method and system based on score in predicting and user characteristics | |
CN110111184A (en) | A kind of negative sequence recommended method and system based on weighting Bayesian inference |
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 |
Application publication date: 20190820 |
|
RJ01 | Rejection of invention patent application after publication |