CN110490663A

CN110490663A - A kind of data processing method, device and electronic equipment

Info

Publication number: CN110490663A
Application number: CN201910786369.2A
Authority: CN
Inventors: 孙艳芳; 陆爱萍
Original assignee: Lenovo Beijing Ltd
Current assignee: Lenovo Beijing Ltd
Priority date: 2019-08-23
Filing date: 2019-08-23
Publication date: 2019-11-22

Abstract

This application discloses a kind of data processing method, device and electronic equipment, method include: acquisition and target object it is associated at least one data are described, the description data are at least one persons of description for target object generation；At least one describes at least one attribute of data described in acquisition is corresponding, described to belong to the person of description described in performance characterization to the satisfied tendency of the target object；At least based at least one described attribute, processing result is obtained, the processing result can characterize the satisfaction of the target object.It can be seen that, it is obtained in the application by the description data to target object, and then description person can be characterized to it, attribute of satisfied tendency of target object is analyzed, it can be obtained by processing result, description person is characterized to the satisfaction of target object with this processing result, so as to reduce the time-consuming in satisfaction acquisition, the efficiency for obtaining satisfaction is thus effectively improved.

Description

A kind of data processing method, device and electronic equipment

Technical field

This application involves technical field of data processing more particularly to a kind of data processing methods, device and electronic equipment.

Background technique

Currently, generalling use the modes such as return visit, questionnaire survey in order to obtain product customer satisfaction and carrying out, cause to be satisfied with Time consumed by the process of degree is longer, so that the efficiency of satisfaction investigation is lower.

Therefore, a kind of technical solution that can be improved and obtain the efficiency of product customer satisfaction is needed.

Summary of the invention

In view of this, the application provides a kind of data processing method, device and electronic equipment, to improve acquisition satisfaction Efficiency.

This application provides a kind of data processing methods, comprising:

Obtain with target object it is associated at least one data are described, the description data are that at least one person of description is directed to What the target object generated；

At least one describes at least one attribute of data, the person of description couple described in the category performance characterization described in acquisition is corresponding The satisfied tendency of the target object；

At least based at least one described attribute, processing result is obtained, the processing result can characterize the target object Satisfaction.

The above method optionally at least based at least one described attribute, obtains processing result, comprising:

The first quantitative value is obtained, first quantitative value is the total quantity of the description data of the attribute characterization first tendency Value；

The second quantitative value is obtained, second quantitative value is the total quantity of the description data of the attribute characterization second tendency Value；Wherein, first tendency is opposite with second tendency；

It is at least based on first quantitative value and second quantitative value, obtains processing result, the processing result includes For the value of the satisfaction of the target object.

At least one describes at least one rank of data described in acquisition is corresponding, and the rank can characterize the description person couple The tendency rank of the target object, wherein the tendency rank is corresponding with level weights value；

Obtain the first quantitative value, first quantitative value are as follows: the corresponding rank of rank described at least one described rank Weighted value corresponding with the rank first describes the sum of products of quantitative value, and the rank corresponding first describes quantitative value and is The total magnitude for the description data that the corresponding attribute characterization first of the rank is inclined to；

Obtain the second quantitative value, second quantitative value are as follows: the corresponding rank of rank described at least one described rank Weighted value corresponding with the rank second describes the sum of products of quantitative value, and the rank corresponding second describes quantitative value and is The total magnitude for the description data that the corresponding attribute characterization second of the rank is inclined to；Wherein, second tendency and institute It is opposite to state the first tendency；

The above method is optionally at least based on first quantitative value and second quantitative value, obtains processing result, Include:

It obtains first quantitative value and subtracts the difference that second quantitative value obtains；

Obtain first quantitative value is obtained plus second quantitative value and value；

Obtain the difference and the ratio between value；

It is at least based on the ratio, obtains the value of the satisfaction for the target object.

The above method is optionally at least based on the ratio, obtains the value of the satisfaction for the target object, packet It includes:

By the ratio multiplied by preset amplification coefficient, the value of the satisfaction for the target object is obtained.

The above method, optionally, at least one describes at least one attribute of data described in the acquisition correspondence, comprising:

Using marking model, to it is described at least one describe data and parse, obtain it is described at least one data are described At least one attribute, the attribute characterization first tendency or the second tendency；

Wherein, the marking model is using at least two there is the sample training of default tendency label to obtain.

At least one describes at least one attribute for target sub-object of data described in acquisition is corresponding, wherein described Belong to the person of description described in performance characterization to be inclined to the satisfied of the target sub-object, the target sub-object is the target object Part；

Wherein, the target sub-object is for enabling the processing result to characterize the description person to the target object Local satisfaction.

The above method, optionally, further includes:

Data are described to described at least one to pre-process；

Wherein, the pretreatment includes: to classify to the description data according to the object identity of the target object.

Present invention also provides a kind of data processing equipments, comprising:

Data acquiring unit, for obtain with target object it is associated at least one data are described, the description data are At least one person of description generates for the target object；

Attribute obtaining unit, for obtain it is corresponding described at least one at least one attribute of data, the attribute are described The description person can be characterized to be inclined to the satisfied of the target object；

As a result obtaining unit, at least based at least one described attribute, obtaining processing result, the processing result energy Characterize the satisfaction of the target object.

Present invention also provides a kind of electronic equipment, comprising:

Memory runs generated data for storing application program and the application program；

Processor, for executing the application program, to realize function: obtain with target object it is associated at least one Data are described, the description data are that at least one person of description generates for the target object；It obtains described in corresponding at least At least one attribute of one description data, it is described to belong to the person of description described in performance characterization to the satisfied tendency of the target object； At least based at least one described attribute, processing result is obtained, the processing result can characterize the satisfaction of the target object.

It can be seen from the above technical proposal that a kind of data processing method, device and electronic equipment disclosed in the present application, In After description person is got for description data caused by target object, target object is expired by obtaining characterization description person The attribute for tendency of anticipating characterizes the satisfaction for target object to obtain processing result with this.As it can be seen that passing through in the application The description data of target object are obtained, and then description person can be characterized to it to the attribute of the satisfied tendency of target object It is analyzed, so that it may obtain processing result, description person be characterized to the satisfaction of target object with this processing result, to subtract The time-consuming that few satisfaction obtains, thus, it is possible to effectively improve the efficiency for obtaining satisfaction.

Detailed description of the invention

Technical solution in ord to more clearly illustrate embodiments of the present application, below will be to required use in embodiment description Attached drawing be briefly described, it should be apparent that, the drawings in the following description are only some examples of the present application, for this For the those of ordinary skill of field, without creative efforts, it can also be obtained according to these attached drawings others Attached drawing.

Fig. 1 is a kind of flow chart for data processing method that the embodiment of the present application one provides；

Fig. 2 and Fig. 3 is respectively a kind of partial process view for data processing method that the embodiment of the present application one provides；

Fig. 4 is a kind of structural schematic diagram for data processing equipment that the embodiment of the present application two provides；

Fig. 5 is the structural schematic diagram for a kind of electronic equipment that the embodiment of the present application three provides.

Specific embodiment

Below in conjunction with the attached drawing in the embodiment of the present application, technical solutions in the embodiments of the present application carries out clear, complete Site preparation description, it is clear that described embodiments are only a part of embodiments of the present application, instead of all the embodiments.It is based on Embodiment in the application, it is obtained by those of ordinary skill in the art without making creative efforts every other Embodiment shall fall in the protection scope of this application.

As shown in Figure 1, a kind of implementation flow chart of the data processing method provided for the embodiment of the present application one, the present embodiment In method can be adapted for being mainly used in the equipment such as the computer for being able to carry out data processing or server to description person's needle The satisfaction of one or certain multiple target object are analyzed.

Specifically, the method in the present embodiment may comprise steps of:

Step 101: obtain with target object it is associated at least one data are described.

Wherein, description data are that at least one person of description generates for target object.For example, description data may include Have: literal expression, picture publication, the use of expression packet and fractional levels label that description person carries out for target object etc. are therein One or more description data.Description person can characterize the table that description person is directed to target object by its description data generated Up to content and Sentiment orientation.Specifically, description data can be one, the corresponding subsequent analysis that carries out can have specific aim；It retouches State data may be it is multiple, corresponding subsequent analysis can have extensive use, it is also possible that analysis result it is more accurate.

It should be noted that target object can be the object that description person is of interest or uses, such as such as some physical product The objects such as notebook, certain Virtual Service such as scene of game or certain application function such as chat application or shopping application.This reality It applies and is intended to analyze satisfaction of the description person to these objects in example.

Specifically, can be obtained and target pair from the various network platforms such as electric business, forum or from media in the present embodiment As associated description data.

Step 102: acquisition corresponds at least one and describes at least one attribute of data.

Wherein, attribute can characterize description person and be inclined to the satisfied of target object.The attribute of data is described in the present embodiment It can be to describe the emotion attribute embodied in data, such as positive emotion or negative sense emotion, or recommend emotion or contradict emotion etc..

Specifically, can be by perhaps punctuation mark etc. in the word in description data according to preset category in the present embodiment Property type carry out similarity classification, to obtain the attribute of corresponding description data, to symbolize satisfaction of the description person to target object Tendency；

Alternatively, can be corresponded to by carrying out the semantic analysis based on deep learning to description data in the present embodiment The attribute of data is described, is inclined to symbolizing description person to the satisfied of target object；

Alternatively, in the present embodiment can by description data corresponding to other information as description the time and/or whether The information such as additional description are analyzed, to obtain the attribute of corresponding description data, to characterize satisfaction of the description person to target object Tendency.

It should be noted that the satisfied tendency that attribute is characterized can be understood as description person to interior provided by target object Whether appearance, function or service etc. are satisfied with or satisfaction etc..

Step 103: being at least based at least one attribute, obtain processing result.

Wherein, processing result can characterize the satisfaction of target object.Specifically, processing result can characterize description person for mesh Mark the satisfaction of object, such as client to the satisfaction of notebook or user to the satisfaction of chat application.

Specifically, can be characterized in the present embodiment by quantifying to the satisfied tendency in attribute with obtaining attribute Satisfied tendency corresponding be satisfied with angle value；Alternatively, can also be carried out according to preset satisfaction rank to attribute in the present embodiment Classification, to obtain the corresponding satisfaction rank of satisfaction that attribute is characterized；Alternatively, can incline to characterization satisfaction in the present embodiment To attribute corresponding to description data it is for statistical analysis, satisfied be inclined to corresponding satisfaction with calculate that attribute characterized Value, etc..

It should be noted that in the present embodiment carry out Analysis of Satisfaction institute according to description data can be a description person It is generated for target object, correspondingly, obtained processing result is the satisfaction for characterizing the person of description and being directed to target object Degree as a result, it is possible thereby to analyze certain specific description persons be directed to target object satisfaction；Alternatively, being carried out in the present embodiment Analysis of Satisfaction institute according to description data also may include multiple description persons be directed to target object generate data, correspondingly, Obtained processing result is the satisfaction that can characterize these persons of description for the target object as a result, it is possible thereby to analyzing Most description persons are directed to the satisfaction of certain specific objective objects out.

By above scheme it is found that a kind of data processing method provided by the embodiments of the present application, is directed to getting description person After description data caused by target object, by obtaining characterization description person to the attribute of the satisfied tendency of target object, come Processing result is obtained, the satisfaction for target object is characterized with this.As it can be seen that by being retouched to target object in the present embodiment It states data to be obtained, and then description person can be characterized to it, attribute of satisfied tendency of target object is analyzed, so that it may To obtain processing result, description person is characterized to the satisfaction of target object with this processing result, is no longer dependent on questionnaire survey Etc. modes, thus reduce satisfaction acquisition time-consuming, thus, it is possible to effectively improve obtain satisfaction efficiency.

Further, in the present embodiment by from electric business, from description data obtained in the modes such as media and/or forum There can be popularity relative to modes such as questionnaire surveys, be retouched as a result, by expanding for what satisfaction obtained in the present embodiment Source or the range of data are stated, so that subsequent obtained satisfaction is more accurate.

It should be noted that the technical solution in the present embodiment can be adapted for the Satisfaction index for a target object Analysis, describes data for a target object at this time；Technical solution in the present embodiment is readily applicable to for multiple The Analysis of Satisfaction of target object, describes data for multiple and different target objects at this time, by description data according to After object is classified, the scheme in the present embodiment is executed for each target object respectively, is achieved in each target The Analysis of Satisfaction of object.

Specifically, can be pre-processed to description data, in the present embodiment after obtaining description data for example, right Description data are classified according to the object identity of target object.Since description data may be from different platform or service Device, correspondingly, the title of target object of the description data corresponding to separate sources or form of presentation be not identical, such as same production Used statement word may be different on different shopping platforms for product, but the case where really belong to identical product, at this point, this The accuracy rate and efficiency of data processing are further increased in embodiment, first can will before carrying out data processing Description data are sorted out according to the object identity of target object, the description data for belonging to the same target object are grouped into same In class, it thus can have and Analysis of Satisfaction targetedly carried out respectively to target object.

In one implementation, the step 102 in the present embodiment, specifically can be in the following manner when obtaining attribute It realizes:

Using marking models, data are described at least one and are parsed, obtain at least one describe data at least one A attribute.

Wherein, the tendency of attribute characterization first or the second tendency, the first tendency refer to that description person is directed to the emotion of target object Tendency, such as positive Sentiment orientation, for example, user likes or evaluate positive tendency to chat application, the second tendency refers to Description person is directed to the tendency opposite with the first tendency of target object, such as the Sentiment orientation of negative sense, for example, client detests service Dislike or evaluate poor tendency.That is, carrying out tendency mark to description data using marking model in the present embodiment, in turn The attribute of corresponding description data is obtained, for example, being labeled as positive or negative sense Sentiment orientation to some verbal description data.

It in oneainstance, may include a tendency description content in a description data, which only describes Satisfied tendency in target object entirety or some dimension (can be understood as a subobject in target object), such as " pen Remember that this shell feel is fine ", at this point, carrying out attribute labeling to the tendency description content using marking model, an attribute is obtained, For example, the attribute of description data is noted as positive Sentiment orientation (for target object entirety or the positive emotion of subobject Tendency)；

In another scenario, the tendency description content there are two may include in a description data or more, often A tendency description content describes the satisfied tendency in the several dimensions of certain in target object entirety or target object, such as " notes respectively This shell feel is fine, and desktop layouts are also pretty good, but has blocked very much bad ", at this point, using marking model respectively to each tendency Description content carries out attribute labeling, obtains the attribute of each tendency description content, and attribute characterization describes in the corresponding description of data The person of description is inclined to the satisfied of target object in appearance, later, in one implementation, can be by the attribute of each tendency description content The attribute that data are described as this describes data with multiple attributes, alternatively, in another implementation, it can be to these Attribute is handled, and an attribute is generated, as follows as the attribute of description data:

In one implementation, the first tendency or the second tendency that can be characterized according to attribute in the present embodiment, to multiple Attribute carries out counting counteracting, for example, the quantity for the attribute that characterization first is inclined to and characterization second tendency attribute quantity into Row compares, if the quantity of the attribute of the first tendency of characterization is greater than the quantity of the attribute of the second tendency of characterization, to description number According to attribute labeling be characterize first tendency attribute, if characterization second tendency attribute quantity be greater than characterization first tendency Attribute quantity, then to description data attribute labeling be characterize second be inclined to attribute；

Such as " notebook computer shell feel is fine, and desktop layouts are also pretty good, but has blocked very much bad " corresponding three attributes: The negative sense attribute of the positive attribute of shell, the positive attribute of desktop and memory obtains description data after counting counteracting Characterize positive attribute；

In another implementation, the first tendency or the second tendency that can be characterized according to attribute in the present embodiment, to more A attribute is weighted counteracting, for example, the attribute weight of different description contents is preset, using these attribute weights to table The quantity of the attribute of the first tendency of sign is weighted summation, i.e., the quantity for the attribute being inclined to each characterization first respectively with the category Property corresponding attribute weight be multiplied after sum it up, and by the quantity of the attribute of each the second tendency of characterization respectively with the attribute Corresponding attribute weight sums it up after being multiplied, and is finally compared two and value, according to corresponding to biggish and value It is inclined to mark the attribute of description data, i.e., is biggish and the corresponding attribute of value tendency by the attribute labeling for describing data. Such as " notebook computer shell feel is fine, and desktop layouts are also pretty good, but has blocked very much bad " corresponding three attributes: shell is just To attribute, the negative sense attribute of the positive attribute of desktop and memory, wherein shell attribute weight is 0.1, and desktop attribute weight is 0.2, memory attribute weight 0.5 is weighted summation and compares with value, determines the attribute of the characterization negative sense of description data.

Specifically, the marking model in the present embodiment is using at least two there is the sample training of default tendency label to obtain It arrives.Wherein, the sample of marking model can be the description data of history, if at least one person of description is at least one target object Generated multiple description samples have default tendency label in each description sample, for example, sample 1 and sample 2 have forward direction Label, sample 3, sample 4 and sample 5 have negative sense label, and sample 6 has positive label, correspondingly, utilizing in the present embodiment These samples are trained, to obtain marking model, marking model can be to not based on the training of sample and its tendency label Know that the description data of tendency are labeled, to obtain the attribute of the description data of unknown tendency, so that description data have accordingly Sentiment orientation, such as positively or negatively.

It should be noted that the marking model in the present embodiment can be the machine learning based on natural language processing technique Network model first to the word or character progress word segmentation processing in description data, then based on sample and its inclines in marking model Training to label carries out Emotion tagging to the word section or character string that obtain by participle, so that marking model was exported As a result include in: description data belong to the probability value of every kind of tendency, and then choose the biggish Sentiment orientation of probability value to generate This describes the attribute of data, and the attribute characterization person of description is that the biggish emotion of probability value chosen is inclined to the tendency of target object at this time To, such as the second tendency of the first positive tendency or negative sense.

In the concrete realization, attribute obtained in step 102 can be in a broad sense for the entirety of target object The attribute of satisfied tendency is also possible to the attribute for certain local satisfied tendencies in target object.For example, to notebook The satisfied tendency attribute of the forward direction of product or the attribute of the satisfied tendency of negative sense as a whole, or to the display in notebook Or the satisfied tendency attribute of forward direction or the satisfied attribute being inclined to of negative sense of keyboard.The part of target object is it is to be understood that target pair Local part or aspect as in, as on notebook keyboard or display, speed of service or image quality in game application etc..

Description data are parsed correspondingly, can use the modes such as marking model in the present embodiment in a step 102, On the level for the total satisfactory grade of target object, the characterization description person of corresponding description data is obtained to target object entirety Satisfied tendency attribute, to obtain characterization to the processing knot of target object total satisfactory grade on the basis of these attributes Fruit；

Description data are parsed alternatively, can use the modes such as marking model in the present embodiment in a step 102, In The characterization description person for describing data is corresponded to some in target object for obtaining in the level of the local satisfaction of target object The attribute of the satisfied tendency of target sub-object or certain multiple target sub-object, to be characterized on the basis of these attributes To the processing result of the local satisfaction in target object.At this point, target sub-object can integrally be satisfied with characterization target object The processing result of degree is refined, so that finally obtained processing result is the part satisfaction for characterizing description person for target object Degree (satisfaction of such as target sub-object) as a result, be achieved in refinement of the satisfaction on different dimensions, such as pass through some Or the determination of multiple product parts, it is refined on product total satisfactory grade, obtains user to product on different product part Satisfaction.

Specifically, step 102 can be accomplished by the following way:

Firstly, carrying out data screening for target sub-object to description data, will describe will be with target sub-object in data Associated data are chosen, and reject data unrelated with target sub-object in description data, and thereby, it is possible to reduce data The amount of processing, so as to accelerate efficiency；

Later, it analyzes by the description data of data screening, specifically can use marking model to by screening Description data be labeled, description person can be characterized for the category of the satisfied tendency of target sub-object with obtain description data Property.

In subsequent, then based on these it is labeled with the attribute in the description data of attribute, to obtain processing result, handled at this time As a result local satisfaction of the characterization description person to target object, that is to say, that processing result at this time is that can characterize description person couple The result of the satisfaction of target sub-object, it is seen then that target sub-object is determined to refine to so that processing result is refined Description person can be characterized to the part i.e. satisfaction of target sub-object of target object.

Wherein, marking model here can have and incline to be obtained using multiple sample trainings with tendency label To the sample of label can be it is all have the relevant sample of target object, be also possible to only with the target sub-object in target object Relevant sample, correspondingly, trained marking model can to description data in characterization description person it is right to target The attribute of the satisfied tendency of elephant is labeled.

Alternatively, step 102 can also be accomplished by the following way:

Firstly, analyzing description data, it specifically can use marking model and description data be labeled, to obtain Description data can characterize description person for the attribute of the satisfied tendency of target object；

Later, the description data for having marked attribute are screened, is rejected unrelated with target sub-object in description data Data, only will description data in data relevant to target sub-object choose, the category of the description data selected at this time Property characterization description person the satisfied of target sub-object is inclined to.

For example, can be first after carrying out attribute labeling to be described data, in the description that attribute characterization first is inclined to Data relevant to target sub-object are filtered out in data, are filtered out and target in the description data that attribute characterization second is inclined to The relevant data of subobject；Alternatively, can be regardless of tendency, after be described data carry out attribute labeling, only according to description Whether data are related to target sub-object to be screened, and the attribute characterization person of description of the description data selected is to target sub-object Satisfied tendency.

In subsequent, then based on filter out these be labeled with the attribute in the description data of attribute, to obtain processing result, Local satisfaction of the processing result characterization description person to target object at this time, that is to say, that processing result at this time is that can characterize Result of the description person to the satisfaction of target sub-object, it is seen then that target sub-object is determined to so that processing result carries out carefully Change, refine to the part i.e. satisfaction of target sub-object that can characterize description person to target object.

In one implementation, step 103 can be accomplished by the following way in the present embodiment, as shown in Figure 2:

Step 201: obtaining the first quantitative value.

Wherein, the first quantitative value is the total magnitude for the description data that attribute characterization first is inclined to, for example, client is to product Evaluation data in positive evaluation data total amount of data, or if user is to marking positive in the marking data of chat application The total amount of data of data, customer evaluates the expression packet of shopping application the total amount of data of expression packet positive in data for another example, etc. Deng.

Step 202: obtaining the second quantitative value.

Wherein, the second quantitative value is the total magnitude for the description data that attribute characterization second is inclined to.For example, client is to product Evaluation data in negative sense evaluation data total amount of data, or such as marking of the user to negative sense in the marking data of chat application The total amount of data of data, customer evaluates the expression packet of shopping application the total amount of data of the expression packet of negative sense in data for another example, etc. Deng.

Step 203: being at least based on the first quantitative value and second quantitative value, obtain processing result.

Wherein, in processing result include value for the satisfaction of target object.

Specifically, first quantitative value can be tied divided by the obtained ratio of the second quantitative value as processing in step 203 For the value of the satisfaction of target object in fruit, alternatively, further, it can be by the first processing costs divided by the ratio of the second quantitative value Value is multiplied by the value after preset amplification coefficient as the satisfaction for being directed to target object in processing result；

Alternatively, the variance of the first quantitative value and the second quantitative value can be obtained in step 203 as being directed in processing result The value of the satisfaction of target object, alternatively, further, it can be on the basis of the variance of the first quantitative value and the second quantitative value Multiplied by amplification coefficient, using obtained product as the value for the satisfaction for being directed to target object in processing result；

Alternatively, can also be accomplished by the following way in step 203:

Firstly, obtaining the first quantitative value subtracts the obtained difference of the second quantitative value, then the first quantitative value is obtained plus the Two quantitative values are obtained and are worth；Later, difference and the ratio between value are obtained, is finally at least directed to based on ratio The value of the satisfaction of target object.

For example, obtained product is made multiplied by a preset amplification coefficient, such as 10 or 100 on the basis of ratio For the value of the satisfaction for target object, the gap between the value of satisfaction numerically is pulled open from there through amplification coefficient, So that can more intuitively be compared in the effect of amplification coefficient between the value of satisfaction, for example, being satisfied with angle value for 0.5 or 0.8 Multiplied by 10 amplification coefficient, 5 or 8 are obtained, it can more intuitive progress satisfaction measurement；

Alternatively, subtract corresponding a reference value in ratio, then the difference after a reference value will be subtracted multiplied by amplification coefficient, this When, obtained product is re-used as the value of the satisfaction for target object, subtracts base in ratio first in the present embodiment as a result, Quasi- value amplifies again, so that the gap between satisfaction is further pulled open, so that can be in base between the value of satisfaction The effect of quasi- value and amplification coefficient more intuitively compares, for example, by 0.5 or 0.8 be satisfied with angle value be individually subtracted 0.3 and 0.2 it Afterwards, multiplied by 10 amplification coefficient, obtain 2 or 6, can more intuitive progress satisfaction measurement, etc..

In another implementation, step 103 can be accomplished by the following way in the present embodiment, as shown in Figure 3:

Step 301: acquisition corresponds at least one and describes at least one rank of data.

Wherein, rank can characterize description person to the tendency rank of target object, for example, emotion of the description person to target object Tendency includes that there are many be inclined to rank: first, second, third, fourth and fifth rank, etc., even more, every kind of tendency grade Does not successively decrease or be incremented by with this in Sentiment orientation degree, such as most strong, tendency grade strongly, stronger, common and noninductive Not, etc..

By taking user is to the hobby of service as an example, forward direction is inclined to corresponding five kinds of tendency ranks: super to like, enjoy a lot, comparing It prefers, commonly like and noninductive five kinds of ranks；Negative sense is inclined to corresponding five kinds of tendency ranks: super to dislike, dislike very much, compare Disagreeable, common disagreeable and noninductive five kinds of ranks, at this point, user be to can be divided into nine kinds of tendency ranks in the hobby description of service, Such as: super to like, enjoy a lot, preferring, commonly liking, is noninductive, commonly disliking, comparing disagreeable, very disagreeable, super beg for Detest.

It should be noted that each tendency rank is corresponding with level weights.Wherein level weights characterization tendency rank is full Significance level in the analysis of meaning degree, the corresponding level weights of all tendency ranks and be 1.The level weights can be according to demand It presets, for example, the level weights that most strong tendency rank is arranged are 0.5, the grade of tendency rank strongly is set Other weight is 0.2, and the level weights that stronger tendency rank is arranged are 0.15, and the level weights of common tendency rank are arranged It is 0.1, the level weights that noninductive tendency rank is arranged are 0.05, etc..

Step 302: obtaining the first quantitative value.

Wherein, the first quantitative value is that level weights value corresponding to each rank is corresponding with the rank at least one rank First sum of products for describing quantitative value, and it is the corresponding category performance characterization of rank that rank corresponding first, which describes quantitative value, The total magnitude of the description data of one tendency.That is, all ranks involved in be described data in the present embodiment In each rank corresponding level weights respectively multiplied by the description data of properties the first tendency of characterization of each rank Total amount and then adduction obtain the first quantitative value.

For example, in evaluation data of the client to product, by the description of the attribute characterization forward direction tendency under most strong rank Attribute characterization forward direction under strongly rank is inclined to by the total magnitude of data multiplied by the level weights of the most strong rank The total magnitude of data is described multiplied by the level weights of the strongly rank, the attribute characterization forward direction under stronger rank is inclined To description data total magnitude multiplied by the level weights of the stronger rank, the attribute characterization forward direction under common grade is inclined To description data total magnitude multiplied by the common grade level weights, by under noninductive rank attribute characterization forward direction be inclined to Description data total magnitude multiplied by the noninductive rank level weights, finally by above 5 products sum it up, obtain first number Magnitude.

Step 303: obtaining the second quantitative value.

Wherein, the second quantitative value is that the corresponding level weights value of rank corresponding with rank second is retouched at least one rank The sum of products of quantitative value is stated, and it is retouching for the corresponding tendency of attribute characterization second of rank that rank corresponding second, which describes quantitative value, State the total magnitude of data.That is, each grade in the present embodiment in all ranks involved in be described data Not corresponding level weights respectively multiplied by each rank it is properties characterization second tendency description data total amount and then Adduction, obtains the second quantitative value.

For example, in evaluation data of the client to product, by the description of the attribute characterization negative sense tendency under most strong rank Attribute characterization negative sense under strongly rank is inclined to by the total magnitude of data multiplied by the level weights of the most strong rank The total magnitude of data is described multiplied by the level weights of the strongly rank, the attribute characterization negative sense under stronger rank is inclined To description data total magnitude multiplied by the level weights of the stronger rank, the attribute characterization negative sense under common grade is inclined To description data total magnitude multiplied by the common grade level weights, by under noninductive rank attribute characterization negative sense be inclined to Description data total magnitude multiplied by the noninductive rank level weights, finally by above 5 products sum it up, obtain second number Magnitude.

Step 304: being at least based on the first quantitative value and second quantitative value, obtain processing result.

Specifically, first quantitative value can be tied divided by the obtained ratio of the second quantitative value as processing in step 304 For the value of the satisfaction of target object in fruit, alternatively, further, it can be by the first processing costs divided by the ratio of the second quantitative value Value is multiplied by the value after preset amplification coefficient as the satisfaction for being directed to target object in processing result；

Alternatively, the variance of the first quantitative value and the second quantitative value can be obtained in step 304 as being directed in processing result The value of the satisfaction of target object, alternatively, further, it can be on the basis of the variance of the first quantitative value and the second quantitative value Multiplied by amplification coefficient, using obtained product as the value for the satisfaction for being directed to target object in processing result；

Alternatively, can also be accomplished by the following way in step 304:

For example, multiplied by a preset amplification coefficient on the basis of ratio, using obtained product as target pair The value of the satisfaction of elephant pulls open the gap between the value of satisfaction from there through amplification coefficient, so that energy between the value of satisfaction Enough effects in amplification coefficient more intuitively compare；

Alternatively, subtract corresponding a reference value in ratio, then the difference after a reference value will be subtracted multiplied by amplification coefficient, this When, obtained product is re-used as the value of the satisfaction for target object, subtracts base in ratio first in the present embodiment as a result, Quasi- value amplifies again, so that the gap between satisfaction is further pulled open, so that can be in base between the value of satisfaction The effect of quasi- value and amplification coefficient more intuitively compares.

With reference to Fig. 4, for a kind of structural schematic diagram for data processing equipment that the embodiment of the present application two provides, which can be with Suitable for the equipment such as computer or server, specifically, the device in the present embodiment may include with flowering structure:

Data acquiring unit 401, for obtain with target object it is associated at least one data are described.

Attribute obtaining unit 402, for obtain it is corresponding described at least one at least one attribute of data is described.

Alternatively, the category of corresponding description data can be obtained in the present embodiment by carrying out deep learning to description data Property, it is inclined to symbolizing description person to the satisfied of target object；

As a result obtaining unit 403, at least based at least one described attribute, obtaining processing result.

By above scheme it is found that a kind of data processing equipment provided by the embodiments of the present application, is directed to getting description person After description data caused by target object, by obtaining characterization description person to the attribute of the satisfied tendency of target object, come Processing result is obtained, the satisfaction for target object is characterized with this.As it can be seen that by being retouched to target object in the present embodiment It states data to be obtained, and then description person can be characterized to it, attribute of satisfied tendency of target object is analyzed, so that it may To obtain processing result, description person is characterized to the satisfaction of target object with this processing result, is no longer dependent on questionnaire survey Etc. modes, thus reduce satisfaction acquisition time-consuming, thus, it is possible to effectively improve obtain satisfaction efficiency.

In one implementation, as a result obtaining unit 403 is specifically used for:

The first quantitative value is obtained, first quantitative value is the total quantity of the description data of the attribute characterization first tendency Value；The second quantitative value is obtained, second quantitative value is the total magnitude of the description data of the attribute characterization second tendency；Its In, first tendency is opposite with second tendency；It is at least based on first quantitative value and second quantitative value, is obtained Processing result, the processing result include the value for the satisfaction of the target object.

Alternatively, result obtaining unit 403 is specifically used for:

At least one describes at least one rank of data described in acquisition is corresponding, and the rank can characterize the description person couple The tendency rank of the target object, wherein the tendency rank is corresponding with level weights value；The first quantitative value is obtained, it is described First quantitative value are as follows: the corresponding level weights value of rank described at least one described rank corresponding with the rank first is retouched State the sum of products of quantitative value, it is the corresponding attribute characterization of the rank that the rank corresponding first, which describes quantitative value, The total magnitude of the description data of one tendency；Obtain the second quantitative value, second quantitative value are as follows: at least one described rank The corresponding level weights value corresponding with the rank second of the rank describes the sum of products of quantitative value, and the rank is corresponding Second the total magnitude that quantitative value is the description data that the corresponding attribute characterization second of the rank is inclined to is described；Its In, second tendency and first tendency are opposite；It is at least based on first quantitative value and second quantitative value, is obtained Processing result, the processing result include the value for the satisfaction of the target object.

Wherein, as a result obtaining unit 403 is at least being based on first quantitative value and second quantitative value, is handled When as a result, it can be accomplished by the following way:

It obtains first quantitative value and subtracts the difference that second quantitative value obtains；First quantitative value is obtained to add Second quantitative value obtains and value；Obtain the difference and the ratio between value；It is at least based on the ratio, is obtained The value of the satisfaction for the target object is obtained, for example, obtaining by the ratio multiplied by preset amplification coefficient for described The value of the satisfaction of target object.

In one implementation, attribute obtaining unit 402 is specifically used for: utilize marking model, to it is described at least one Description data are parsed, obtain it is described at least one at least one attribute of data is described, the attribute characterization first is inclined to Or second tendency；Wherein, the marking model is using at least two there is the sample training of default tendency label to obtain.

In another implementation, attribute obtaining unit 402 is specifically used for: at least one describes data described in acquisition is corresponding For at least one attribute of target sub-object, wherein the person of description described in the category performance characterization is to the target sub-object Satisfied tendency, the target sub-object are the part of the target object；

In addition, data acquiring unit 401 first can also describe data to described at least one and be located in advance in the present embodiment Reason；Wherein, the pretreatment includes: to classify to the description data according to the object identity of the target object.

It should be noted that the specific implementation of each functional unit can refer to hereinbefore in the device of the present embodiment Corresponding description, details are not described herein again.

With reference to Fig. 5, for the structural schematic diagram for a kind of electronic equipment that the embodiment of the present application three provides, which can be with The equipment for being able to carry out data processing for computer or server etc..

In the concrete realization, electronic equipment may include with flowering structure:

Memory 501 runs generated data for storing application program and the application program；

Processor 502, for executing the application program, to realize function: obtaining associated at least with target object One description data, the description data are that at least one person of description generates for the target object；Described in being corresponded to At least one describes at least one attribute of data, and the person of description described in the category performance characterization inclines to the satisfaction of the target object To；At least based at least one described attribute, processing result is obtained, the processing result can characterize the satisfaction of the target object Degree.

By above scheme it is found that a kind of electronic equipment provided by the embodiments of the present application, is getting description person for target After description data caused by object, by obtaining characterization description person to the attribute of the satisfied tendency of target object, to obtain Processing result characterizes the satisfaction for target object with this.As it can be seen that passing through the description number to target object in the present embodiment According to being obtained, and then description person can be characterized to it, attribute of satisfied tendency of target object is analyzed, so that it may To processing as a result, characterizing description person to the satisfaction of target object with this processing result, it is no longer dependent on the side such as questionnaire survey Formula, to reduce the time-consuming of satisfaction acquisition, thus, it is possible to effectively improve the efficiency for obtaining satisfaction.

In one implementation, processor 502 is at least at least one attribute based on described in, when obtaining processing result, It is realized especially by following manner:

Alternatively, processor 502 is at least being based at least one described attribute, when obtaining processing result, especially by following Mode is realized:

Wherein, processor 502 is at least being based on first quantitative value and second quantitative value, obtains processing result When, it can be accomplished by the following way:

In one implementation, processor 502 can be accomplished by the following way when obtaining attribute:

Using marking model, to it is described at least one describe data and parse, obtain it is described at least one data are described At least one attribute, the attribute characterization first tendency or the second tendency；Wherein, the marking model is to utilize at least two Sample training with default tendency label obtains.

In another implementation, processor 502 can also be accomplished by the following way when obtaining attribute:

In addition, processor 502 first can describe number to described at least one after obtaining description data in the present embodiment According to being pre-processed；Wherein, the pretreatment includes: to carry out to the description data according to the object identity of the target object Classification.

It should be noted that in the present embodiment processor 502 specific implementation can with reference to hereinbefore accordingly retouch It states, details are not described herein again.

Below to the Analysis of Satisfaction of product in order to, the technical solution in the present embodiment is illustrated, it is as follows:

Firstly, in principle, technical solution in the present embodiment is based on big data processing technique and intelligent algorithm In conjunction with, to user about product comment data carry out orientation of emotion (tendency) identification, by after identification result progress mathematics It calculates, forms measurement index: product customer satisfaction index PSI (Product SatisfactionIndex), the meter of the index It is as follows to calculate formula:

PSI=(quantity of forward direction evaluation data-negative sense evaluation data quantity)/(quantity+negative sense of forward direction evaluation data Evaluate the quantity of data) * index amplification coefficient.

Wherein, the quantity of positive evaluation data may further give difference according to the difference of emotional intensity (tendency rank) Weight coefficient, then again the positive evaluation quantity of the varying strength multiplied by weight coefficient is added to obtain positive evaluation data Quantity.Similarly, the quantity of negative sense evaluation data can also give different weight coefficients according to the difference of emotional intensity, then again The negative sense evaluation quantity of varying strength multiplied by weight coefficient is added to obtain the quantity of negative sense evaluation data.

In addition, index amplification coefficient is arranged according to business demand, wherein consider index can identification suggestion be set as 10 Or 100 equal numerical value.

Specific implementing procedure is as follows:

Firstly, collecting the comment data of client, comment number can be specifically crawled from electric business, forum or from the platforms such as media According to；

Secondly, comment data is stored, meanwhile, comment data can be cleaned and be aligned (according to product mark Knowledge classify) etc. processing；

Later, Emotion tagging is done to comment data with the model etc. of algorithm or building on natural language processing technique (such as Sentiment orientation and emotional intensity mark positively or negatively) and product attribute extract；

Finally, calculating product customer satisfaction index PSI according to calculation formula, wherein PSI can be whole (preceding for product Target object in text) Satisfaction Index, be also possible to for some product attributes (target hereinbefore extracted Object) Satisfaction Index.

Wherein, the whole PSI index of product=(negative sense comment number of the positive comment data quantity-product of the product Data bulk)/(the negative sense comment data quantity of the positive comment data quantity+product of the product) * 10；

The PSI index of some attribute of product=(certain of the positive comment data quantity-product of certain attribute of the product The negative sense comment data quantity of attribute)/(certain attribute of the positive comment data quantity+product of certain attribute of the product it is negative To comment data quantity) * 10.

Wherein, positive comment data quantity can also give weight according to emotional intensity, at this time positive comment data quantity Calculation formula is as follows:

Wherein, i=1,2 ... n, n are the rank sum of positive emotional intensity；

Positive comment data quantity can also give weight according to emotional intensity, and positive comment data quantity calculates public at this time Formula is as follows:

As it can be seen that the technical solution in the present embodiment, simple for the definition of Satisfaction index, and technically can be achieved, and And the consistency that can be realized lasting customer satisfaction is commented.Further, it can be realized customer requirement analysis in the present embodiment Quantitative analysis model, the flexible analysis for supporting various dimensions multi-level.

Each embodiment in this specification is described in a progressive manner, the highlights of each of the examples are with other The difference of embodiment, the same or similar parts in each embodiment may refer to each other.For device disclosed in embodiment For, since it is corresponded to the methods disclosed in the examples, so being described relatively simple, related place is said referring to method part It is bright.

Professional further appreciates that, unit described in conjunction with the examples disclosed in the embodiments of the present disclosure And algorithm steps, can be realized with electronic hardware, computer software, or a combination of the two, in order to clearly demonstrate hardware and The interchangeability of software generally describes each exemplary composition and step according to function in the above description.These Function is implemented in hardware or software actually, the specific application and design constraint depending on technical solution.Profession Technical staff can use different methods to achieve the described function each specific application, but this realization is not answered Think beyond scope of the present application.

The step of method described in conjunction with the examples disclosed in this document or algorithm, can directly be held with hardware, processor The combination of capable software module or the two is implemented.Software module can be placed in random access memory (RAM), memory, read-only deposit Reservoir (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technology In any other form of storage medium well known in field.

The foregoing description of the disclosed embodiments makes professional and technical personnel in the field can be realized or use the application. Various modifications to these embodiments will be readily apparent to those skilled in the art, as defined herein General Principle can be realized in other embodiments without departing from the spirit or scope of the application.Therefore, the application It is not intended to be limited to the embodiments shown herein, and is to fit to and the principles and novel features disclosed herein phase one The widest scope of cause.

Claims

1. a kind of data processing method, comprising:

Obtain with target object it is associated at least one data are described, the description data are described at least one person of description is directed to What target object generated；

At least one describes at least one attribute of data described in acquisition is corresponding, and the person of description described in the category performance characterization is to described The satisfied tendency of target object；

At least based at least one described attribute, processing result is obtained, the processing result can characterize expiring for the target object Meaning degree.

2. according to the method described in claim 1, at least obtaining processing result based at least one described attribute, comprising:

The first quantitative value is obtained, first quantitative value is the total magnitude of the description data of the attribute characterization first tendency；

The second quantitative value is obtained, second quantitative value is the total magnitude of the description data of the attribute characterization second tendency； Wherein, first tendency is opposite with second tendency；

It is at least based on first quantitative value and second quantitative value, obtains processing result, the processing result includes being directed to The value of the satisfaction of the target object.

3. according to the method described in claim 1, at least obtaining processing result based at least one described attribute, comprising:

At least one describes at least one rank of data described in acquisition is corresponding, and the rank can characterize the description person to described The tendency rank of target object, wherein the tendency rank is corresponding with level weights value；

Obtain the first quantitative value, first quantitative value are as follows: the corresponding level weights of rank described at least one described rank Value corresponding with the rank first describes the sum of products of quantitative value, and it is described that the rank corresponding first, which describes quantitative value, The total magnitude for the description data that the corresponding attribute characterization first of rank is inclined to；

Obtain the second quantitative value, second quantitative value are as follows: the corresponding level weights of rank described at least one described rank Value corresponding with the rank second describes the sum of products of quantitative value, and it is described that the rank corresponding second, which describes quantitative value, The total magnitude for the description data that the corresponding attribute characterization second of rank is inclined to；Wherein, second tendency and described the One tendency is opposite；

4. according to the method in claim 2 or 3, being at least based on first quantitative value and second quantitative value, obtain Processing result, comprising:

Obtain the difference and the ratio between value；

5. obtaining the satisfaction for being directed to the target object according to the method described in claim 4, being at least based on the ratio Value, comprising:

6. method according to claim 1,2 or 3, the acquisition it is corresponding it is described at least one at least one of data is described Attribute, comprising:

Using marking model, to it is described at least one describe data and parse, obtain it is described at least one data are described extremely A few attribute, the tendency of attribute characterization first or the second tendency；

7. method according to claim 1,2 or 3, the acquisition it is corresponding it is described at least one at least one of data is described Attribute, comprising:

At least one describes at least one attribute for target sub-object of data described in acquisition is corresponding, wherein the attribute The description person can be characterized to be inclined to the satisfied of the target sub-object, the target sub-object is the office of the target object Portion；

Wherein, the target sub-object is for enabling the processing result to characterize the description person to the office of the target object Portion's satisfaction.

8. method according to claim 1,2 or 3, further includes:

Data are described to described at least one to pre-process；

9. a kind of data processing equipment, comprising:

Data acquiring unit, for obtain with target object it is associated at least one data are described, the description data are at least One description person generates for the target object；

Attribute obtaining unit, for obtain it is corresponding described at least one at least one attribute of data, the attribute energy table are described The description person is levied to be inclined to the satisfied of the target object；

As a result obtaining unit, at least based at least one described attribute, obtaining processing result, the processing result can characterize The satisfaction of the target object.

10. a kind of electronic equipment, comprising:

Processor, for executing the application program, to realize function: obtaining and at least one associated description of target object Data, the description data are that at least one person of description generates for the target object；Obtain at least one described in corresponding to At least one attribute of data is described, it is described to belong to the person of description described in performance characterization to the satisfied tendency of the target object；At least Based at least one described attribute, processing result is obtained, the processing result can characterize the satisfaction of the target object.