CN117252656A - Product recommendation method and device and electronic equipment - Google Patents
Product recommendation method and device and electronic equipment Download PDFInfo
- Publication number
- CN117252656A CN117252656A CN202210636020.2A CN202210636020A CN117252656A CN 117252656 A CN117252656 A CN 117252656A CN 202210636020 A CN202210636020 A CN 202210636020A CN 117252656 A CN117252656 A CN 117252656A
- Authority
- CN
- China
- Prior art keywords
- attribute information
- target
- user
- attribute
- product
- 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 62
- 238000012545 processing Methods 0.000 claims abstract description 17
- 238000010606 normalization Methods 0.000 claims abstract description 9
- 238000007637 random forest analysis Methods 0.000 claims description 44
- 238000003066 decision tree Methods 0.000 claims description 34
- 238000012549 training Methods 0.000 claims description 17
- 238000003860 storage Methods 0.000 claims description 16
- 238000004590 computer program Methods 0.000 claims description 15
- 238000004891 communication Methods 0.000 claims description 14
- 238000005070 sampling Methods 0.000 claims description 4
- 238000010586 diagram Methods 0.000 description 10
- 230000008569 process Effects 0.000 description 8
- 230000006399 behavior Effects 0.000 description 7
- 238000009826 distribution Methods 0.000 description 7
- 238000012360 testing method Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 6
- 230000009286 beneficial effect Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 238000000547 structure data Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000007476 Maximum Likelihood Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 230000002085 persistent effect Effects 0.000 description 1
- 238000013138 pruning Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000001052 transient 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
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Development Economics (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Data Mining & Analysis (AREA)
- Entrepreneurship & Innovation (AREA)
- General Engineering & Computer Science (AREA)
- Game Theory and Decision Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the application provides a product recommendation method, a device and electronic equipment, wherein the method comprises the following steps: under the condition that target attribute information of a target user is obtained, determining a first target attribute number of the target attribute information according to a corresponding relation between preset user attribute information and a basic attribute number; determining a target attribute set corresponding to the first target attribute number based on the first target attribute number, wherein the target attribute set comprises at least part of user attribute information with the association degree larger than a threshold value in the target attribute information, and carrying out normalization processing on the at least part of user attribute information to obtain feature attribute information corresponding to the at least part of user attribute information; inputting the characteristic attribute information into a trained prediction model to obtain a product prediction result corresponding to the target user; and recommending the product corresponding to the product prediction result to the target user according to the product prediction result.
Description
Technical Field
The application relates to the field of big data, in particular to a product recommendation method and device and electronic equipment.
Background
With the rapid development of internet technology, providing personalized services for different users is accepted by more and more enterprises, so that more and more enterprises try to predict the demands of different users by conducting intensive research on the users, thereby recommending products and providing services for the users in a targeted manner.
In some scenes, with the continuous development of mobile communication technology, traffic charge is reduced, users have exceeded call demands for traffic, in order to meet the actual demands of rapid increase of user traffic, a clustering method is often adopted to conduct classification prediction according to historical behavior structure data of users, but the historical behavior structure data of users are adopted to conduct classification prediction, and due to the independence of the historical behavior structure data of each user, actual traffic demands of users are often not predicted correctly, and the accuracy and precision of traffic product recommendation are low.
Disclosure of Invention
The embodiment of the application aims to provide a product recommendation method, device and electronic equipment, and accuracy and precision of flow product recommendation are improved.
In a first aspect, an embodiment of the present application provides a product recommendation method, including: under the condition that target attribute information of a target user is obtained, determining a first target attribute number of the target attribute information according to a corresponding relation between preset user attribute information and a basic attribute number; determining a target attribute set corresponding to the first target attribute number based on the first target attribute number, wherein the target attribute set comprises at least part of user attribute information with the association degree larger than a threshold value in the target attribute information, and carrying out normalization processing on the at least part of user attribute information to obtain feature attribute information corresponding to the at least part of user attribute information; inputting the characteristic attribute information into a trained prediction model to obtain a product prediction result corresponding to the target user; and recommending the product corresponding to the product prediction result to the target user according to the product prediction result.
In a second aspect, an embodiment of the present application provides a product recommendation device, including: the first determining module is used for determining a first target attribute number of the target attribute information according to the corresponding relation between the preset user attribute information and the basic attribute number under the condition that the target attribute information of the target user is acquired; a second determining module, configured to determine, based on the first target attribute number, a target attribute set corresponding to the first target attribute number, where the target attribute set includes at least part of user attribute information in the target attribute information, where a degree of association of the user attribute information is greater than a threshold value, and normalize the at least part of user attribute information to obtain feature attribute information corresponding to the at least part of user attribute information; the input module is used for inputting the characteristic attribute information into the trained prediction model to obtain a product prediction result corresponding to the target user; and the recommending module is used for recommending the product corresponding to the product predicting result to the target user according to the product predicting result.
In a third aspect, embodiments of the present application provide an electronic device including a processor, a communication interface, a memory, and a communication bus; the processor, the communication interface and the memory complete communication with each other through a communication bus; the memory is used for storing a computer program; the processor is configured to execute a program stored in the memory, and implement the steps of the product recommendation method according to the first aspect.
In a fourth aspect, embodiments of the present application provide a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the product recommendation method steps as mentioned in the first aspect.
According to the technical scheme provided by the embodiment of the application, under the condition that the target attribute information of the target user is obtained, the first target attribute number of the target attribute information is determined according to the corresponding relation between the preset user attribute information and the basic attribute number; determining a target attribute set corresponding to the first target attribute number based on the first target attribute number, wherein the target attribute set comprises at least part of user attribute information with the association degree larger than a threshold value in the target attribute information, and carrying out normalization processing on the at least part of user attribute information to obtain feature attribute information corresponding to the at least part of user attribute information; inputting the characteristic attribute information into a trained prediction model to obtain a product prediction result corresponding to the target user; and recommending the product corresponding to the product prediction result to the target user according to the product prediction result. The characteristic attribute information with the association relationship can be used as input data of the prediction model, so that the requirements of users can be accurately predicted, the prediction precision of the prediction model is effectively improved, and the precision and accuracy of flow product recommendation are further improved.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings that are required to be used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments described in the present application, and that other drawings may be obtained according to these drawings without inventive effort to a person skilled in the art.
Fig. 1 is a schematic flow chart of a product recommendation method according to an embodiment of the present application;
fig. 2 is a schematic block diagram of a product recommendation device according to an embodiment of the present disclosure;
fig. 3 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
Detailed Description
The embodiment of the application provides a product recommendation method, a device and electronic equipment, which improve the prediction precision of flow products actually required by users and improve the success rate of recommending the flow products to the users.
In order to better understand the technical solutions in the present application, the following description will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only some embodiments of the present application, not all embodiments. All other embodiments, which can be made by one of ordinary skill in the art based on the embodiments herein without making any inventive effort, shall fall within the scope of the present application.
In the following, a description is given of a product recommendation method provided in the embodiment of the present application, and as illustrated in fig. 1, an exemplary implementation body of the product recommendation method may be a server, where the server may be an independent server or a server cluster formed by a plurality of servers, and the server may be a server capable of recommending a product, and the product recommendation method specifically may include the following steps:
in step S101, when the target attribute information of the target user is obtained, the first target attribute number of the target user is determined according to the preset correspondence between the user attribute information and the basic attribute number.
Specifically, the target user may be a user to be recommended for the flow product, and the flow demand of the user needs to be predicted, and the target attribute information of the target user includes, but is not limited to, basic information and historical query records of the target user. The basic information of the user includes, but is not limited to, the user ID, age, sex, post, department, etc., and the basic information of the user may be stored in a user basic information table as shown in table 1, where the information in the user basic information table is only an example, and the user basic information may also be other types of information, and the embodiment of the present application is not limited herein. And extracting historical query records of the user by using the user ID to construct query behavior result data of the user, wherein the query behavior result data of the user comprises but is not limited to query time, a query module and the like, the query module can adopt numbers to number, the number value corresponding to the query module is stored in a table corresponding to the query module and is stored in a user query history record information table shown in table 2, and the query module can be information queried by the user or an object queried by the user. The information in the query history information table is merely an example, and the user query history information table may also be other types of information, which are not limited herein.
The user attribute information includes, but is not limited to, the basic information and the history query record of the user described in the above embodiment, when the user attribute information is extracted, the user ID may be used as a unique identifier, the basic information of the user includes, but is not limited to, the user ID, the age, the sex, the post, the department and the like, the basic information of the user may be stored in a user basic information table as shown in table 1, where the information in the user basic information table is merely an example, and the user basic information may also be other types of information, and the embodiment of the present application is not limited herein; and extracting historical query records of the user by using the user ID to construct query behavior result data of the user, wherein the query behavior result data of the user comprises but is not limited to query time, a query module and the like, the query module can adopt numbers, the number value corresponding to the query module is stored in a table corresponding to the query module and is stored in a user query history record information table shown in table 2, the information in the user query history record information table is only an example, the user query history record information table can also be other types of information, and the embodiment of the application is not limited herein. Combining the user ID as a connection attribute with the user basic information table and the history query record information table to obtain a comprehensive information table shown in table 3, wherein the user ID, the user basic information and the history query record information in the comprehensive information table correspond to each other, the information in the comprehensive information table is only an example, and the comprehensive information table can also be other types of information.
Table 1 user basic information table
Table 2 user query history information table
Table 3 comprehensive information table
After the target attribute information is obtained, the user attribute information is normalized and quantized, because of the limitation of the random forest classifier, and the non-numerical data needs to be quantized. Non-numeric data including, but not limited to, symbols, status, categories, etc.; the non-numerical data may have two or more states, and the normalized target attribute information corresponds to respective numbers, wherein each attribute information is numbered for the user attribute information as sample data, and there is a correspondence between the attribute information of the user and the basic attribute number, so that after the target attribute information is acquired, the first target attribute number of the target attribute information is determined according to the correspondence. For example, a certain enterprise has M departments, namely, a market department, a human resource department and an administrative department, and in this embodiment of the present application, each department (for example, 1,2,3,4, …, M) may be identified by a set of integers, where these integers are used for data processing, and are not represented by any specific order, and are used for identifying data, for example, the market department is identified by an integer 1. For continuous attributes, integers may be used for identification, for example, ages are classified into elderly, middle aged and young, and it is assumed that younger than 40 years old is identified by integer 1, middle aged 40 to 60 years old is identified by numeral 2, and older than 60 years old is identified by numeral 3. And numbering all the inquiry information modules to be used as the category attribute of each record in the comprehensive information table, wherein the number represents the category attribute. For example, the attribute a1 is numbered with the numeral 1, and the numeral 1 represents the attribute a1.
In step S103, a target attribute set corresponding to the first target attribute number is determined based on the first target attribute number, and the target attribute set includes at least part of user attribute information in which the association degree in the target attribute information is greater than the threshold.
Wherein determining, based on the first target attribute number, a target attribute set corresponding to the first target attribute number includes: and respectively expanding a preset step length to a first direction and a second direction of the attribute number set by taking the first target attribute number as a starting point to obtain a second target attribute number related to the first target attribute number, and determining that user attribute information corresponding to the second target attribute number in the target attribute information belongs to the target attribute set.
Specifically, attribute information of users has a certain degree of association, and the influence of the attribute information of the users with a certain degree of association on the accuracy of the decision tree changes exponentially, and as the number of the attribute information of the users increases, the algorithm complexity of the random forest increases exponentially, so that the decision speed of the random forest is influenced. Therefore, in order to improve the prediction precision of the random forest flow prediction model and reduce the algorithm complexity of the random forest, the attribute information of the user can be replaced by the 'user characteristic attribute information' with a certain degree of association. The first direction may be a left direction in the attribute number set, and the second direction may be a right direction in the attribute number set.
The basic attribute numbers in the attribute number set are the original sample space C to be quantized, which contains the basic attributes m as shown in the first row of table 3, one category attribute (query module), and a plurality of pieces of sample data as shown in the second row of table 3. Given a mutually different sequence for each basic attributeNumber (sequence number is variable), denoted in turn as (1), (2), (i., (m), definition a (m) As the basic attribute index corresponding to the serial number (m), similarly, (1), (2), and the basic attribute index of (m) is defined as a in order (1) ,...,a (m) 。
The central structuring process refers to the basic attribute labeledIs the starting point, the basic attribute of which the step length is k (generally k=1) is synchronously extended in the left and right directions +.>Next, a basic attribute set corresponding to the ω -group basic attribute index shown in the following formula and sample data thereof are obtained, and such a way of constructing the attribute set is called center structuring. Wherein->The function representation is rounded down.
M
Therefore, the central structuring process adopts the mode of marking the serial numbers to arrange attribute information, then uses the attribute corresponding to the central serial numbers as a starting point, expands the attribute of a certain step length, analyzes the characteristics of each subunit attribute, and mines the structured characteristic attribute, thereby effectively improving the classification and prediction precision of the random forest flow prediction model, improving the over-fitting phenomenon caused by resisting data disturbance, reducing the training time of the random forest flow prediction model and improving the training efficiency of the random forest flow prediction model.
In step S105, normalization processing is performed on at least part of the user attribute information, so as to obtain feature attribute information corresponding to at least part of the user attribute information.
Feature attributes refer to a set of basic attributesAnd all quantized data samples thereof, firstly, normalizing the value of each data sample to defineThe feature attributes of this set of basic attributes. Similarly, we can obtain ω feature attributes, denoted S in turn 1 ,S 2 ,...,S ω . Wherein s is (i) Sample values representing characteristic attributes can be obtained, and the sample values of omega characteristic attributes are sequentially expressed as s 1 ,s 2 ,...,s ω 。
In step S107, the feature attribute information is input to the trained prediction model, and a product prediction result corresponding to the target user is obtained.
Specifically, the flow prediction model may be a random forest-based prediction model, and before training the flow prediction model, attribute information of a user serving as a training sample and a test sample is acquired, the attribute information of the user may be used as a data sample for training, testing and optimizing the random forest, and the attribute information of the user may be attribute information of the user acquired from an operator for a certain period of time.
When training the random forest flow prediction model, randomly and repeatedly selecting a first preset number of data from the attribute information of n users as training samples of the random forest flow prediction model, selecting a second preset number of data as test samples of the random forest flow prediction model, wherein the sum of the first preset number and the second preset number is equal to n, the first preset number can be n.times.70%, and the second preset number can be n.times.30%.
By the embodiment, after the training sample set of each decision tree of the random forest flow prediction model is obtained, the random forest flow prediction model is built, and for each decision tree in the random forest, the random decision tree is randomly selected from the training sample setAnd forming a feature subspace by the feature attribute variables to obtain a decision tree. And then calculating the splitting value of each characteristic attribute according to the information gain, and finally taking the optimal result as the splitting criterion of the node. In the classification prediction of random forest, a decision tree is constructed on subspace of each sample set, each decision tree grows without pruning, and finally all decision tree combinations are called as random forest model.
That is, before the random forest traffic prediction model is constructed, the random forest traffic prediction model is initialized, mainly including training samples, test samples, determination of the number of decision trees, and the like of attribute information of the user as in table 3.
In step S109, a product corresponding to the product prediction result is recommended to the target user according to the product prediction result.
Specifically, recommending a product corresponding to the product prediction result to the target user according to the product prediction result includes: and inputting the characteristic attribute information into the trained prediction model, acquiring probability values of prediction results output by each decision tree in the prediction model, ordering the probability values in a descending order, and determining the prediction results ordered in the first N as product prediction results.
Further, in the classification prediction of the random forest traffic prediction model, for the attribute categories in the test samples, for example, in table 3, after all decision trees in the random forest traffic prediction model are determined, the probability that each attribute category is selected is counted, the probability values of the selected attribute categories are sorted in a descending order, and λ (N) attribute categories before being selected as the prediction result according to the user requirements, where λ may be a value according to the actual situation, for example, the value is 1 or 4 or 9 or 16.
According to the technical scheme disclosed by the embodiment of the application, the characteristic attribute information with the association relationship can be used as the input data of the prediction model, so that the demand of a user can be accurately predicted, the prediction precision of the prediction model is effectively improved, and the precision and the accuracy of flow product recommendation are further improved.
In one possible implementation, before inputting the feature attribute information into the trained prediction model, the method further includes: numbering the user attribute information serving as sample data to obtain a first basic attribute number corresponding to the user attribute information;
determining an attribute set corresponding to the first basic attribute number based on the first basic attribute number, wherein the attribute set comprises at least part of user attribute information of which the association degree is greater than a threshold value in the user attribute information, and carrying out normalization processing on at least part of the user attribute information in the attribute set to obtain user characteristic attribute information corresponding to at least part of the user attribute information in the attribute set; sampling data from the user characteristic attribute information based on a target mode to obtain a plurality of sample data sets, wherein each sample data set comprises at least one user characteristic attribute information; constructing a decision tree corresponding to a sample data set, and combining the decision trees to obtain a random forest flow prediction model to be trained; and training the random forest flow prediction model by utilizing the plurality of sample data sets to obtain a trained random forest flow prediction model. Wherein determining the set of attributes corresponding to the first base attribute number based on the first base attribute number includes: and respectively expanding a preset step length to a first direction and a second direction of the attribute number set by taking the first basic attribute number as a starting point to obtain a second basic attribute number related to the first basic attribute number, and determining that user attribute information corresponding to the second basic attribute number belongs to the attribute set.
The following describes an implementation manner of the embodiment of the present application by taking a prediction model as an example of a random forest flow prediction model:
first, the attribute information of the user is subjected to central structuring processing, the characteristic attribute information of the user is constructed, and the constructed characteristic attribute information of the user is shown in table 4.
The target mode may be a Bagging (Bagging) method, and the data sampling is firstly based on the Bagging method. For the random forest flow prediction model, a Bagging method is used for randomly extracting N samples from an original training set X of attribute information of a user (wherein the X comprises N sample data in the form of table 3 in the embodiment) to form a sample set X1, and the method is used for constructing sample sets X2 to X alpha, wherein under the premise of meeting the requirements of the user on the prediction precision and accuracy of the random forest flow prediction model, the smaller the value of alpha is, the better the value of alpha is, and the obtained sample sets X1 to X alpha are used for constructing the training sample set of each decision tree of the random forest flow prediction model, and the number of the training sample sets is the number of the decision trees.
Then, a basic attribute serial number of the sample data, which refers to an integer number of attribute information of the user, is initialized. And then extracting the characteristic attribute based on the central structuring, determining a basic attribute label according to the basic attribute serial number of the sample data, and solving the characteristic attribute of the sample data by adopting a central structuring method. Determining the number of decision trees according to a Bagging method, selecting characteristic attributes and training samples for each decision tree, forming a random forest flow prediction model after all decision trees are built, finally testing the random forest flow prediction model by using a test sample set, determining whether the current random forest flow prediction model is optimal, if not, rearranging the sequence number of attribute information of a user, reinitializing the basic attribute sequence number of sample data, retraining, and if so, determining that the optimal forest is the random forest flow prediction model which is finally trained.
The basic attribute number is that the quantized original sample space C contains m basic attributes as shown in the first row of table 3, one category attribute (query module), and a plurality of pieces of sample data as shown in the second row of table 3. Given a mutually different sequence number (variable) for each basic attribute, denoted in turn as (1),(2) ,., (m), definition a (m) As the basic attribute index corresponding to the serial number (m), similarly, (1), (2), and the basic attribute index of (m) is defined as a in order (1) ,...,a (m) 。
The central structuring process refers to the basic attribute labeledThe basic attribute of (1) is k (generally k=1) which is the starting point and is synchronously extended in the left direction (first direction) and the right direction (second direction)>Next, a basic attribute set corresponding to the ω -group basic attribute index shown in the following formula and sample data thereof are obtained, and such a way of constructing the attribute set is called center structuring. Wherein->The function representation is rounded down.
M
Feature attributes refer to a set of basic attributesAnd all quantized data samples thereof, firstly, normalizing the value of each data sample to defineThe feature attributes of this set of basic attributes. Similarly, we can obtain ω feature attributes, denoted S in turn 1 ,S 2 ,...,S ω . Wherein s is (i) Sample values representing characteristic attributes can be obtained, and the sample values of omega characteristic attributes are sequentially expressed as s 1 ,s 2 ,...,s ω 。
In the embodiment of the application, the feature attribute style obtained by adopting the method is shown in the following table 4.
TABLE 4 Property style sheets
In one possible implementation manner, to further improve the efficiency of the product prediction results output by the prediction model, determining the top N prediction results as the product prediction results includes: under the condition that the predicted results output by all decision trees in the trained predicted model are consistent with the predicted results output by the preset number of decision trees in the predicted model, the predicted results sequenced in the first N are determined to be product predicted results. If the conclusion made by starting h decision trees in the random forest is the same as the conclusion made by all decision trees, the push consistency is considered to be satisfied.
Specifically, for target attribute information of a target user, a certain flow product to be pushed is decided through a forest model. Since the flow product to be pushed has a binary property, if the forest model pushes the current flow product, the sample is called a positive sample and is represented by Rpositive (abbreviated as Rp), and if the forest does not push the flow product, the sample is called a negative sample and is represented by Rnegative (abbreviated as Rn).
For alpha decision trees in a forest, if each classifier (decision tree) independently decides with a probability p whether the sample data is a positive sample, then the probability of h positive samples of the observation results accords with the binomial distribution:
when the number of votes h and the number of trials α of the object are known, the distribution of p is calculated using the Bayes formula.
Where P (p|α) =p (P), which is an information-free prior distribution of P, therefore, by simplifying the formulaThe following steps are obtained:
substituting formula (3) into formula (5) can obtain the binomial joint prior beta distribution as follows:
maximum likelihood estimation of pFor h/α, the β distribution is estimated unbiased as:
in the formula (6), the calculation of the distribution of the random variable p is to guarantee the consistent pushing of the result of the positive sample through probability calculation.
When p is more than or equal to 0.5,the method comprises the following steps:
wherein the method comprises the steps ofFor probability, P (P. Gtoreq.0.5|h, α) is the decision value of the random forest classifier (whether to push the basis of the flow product) if P is in confidence interval [0.5,1]The confidence coefficient of the flow rate is higher than beta (the value of beta depends on the precision required by a user), or lower than 1-beta, the starting of the subsequent classifier can be stopped, and whether the current flow rate product is pushed is determined, so that the decision efficiency of prediction is effectively improved, and the efficiency of the prediction model for outputting the prediction result is further improved. In addition, the reliability prediction of the confidence interval of the beta distribution is adopted, and the corresponding prediction index is obtained only at the cost of a simple lookup table, so that the execution efficiency of the random forest algorithm flow prediction model is greatly improved. In addition, under the condition of higher confidence, the decision accuracy of the random forest flow prediction model is guaranteed.
According to the product recommendation method provided in the foregoing embodiment, based on the same technical concept, the embodiment of the present application further provides a product recommendation device, and fig. 2 is a schematic block diagram of the product recommendation device provided in the embodiment of the present application, where the product recommendation device is configured to execute the product recommendation method described in the foregoing embodiment, and as shown in fig. 2, the product recommendation device 200 includes: a first determining module 201, configured to determine, when target attribute information of a target user is obtained, a first target attribute number of the target attribute information according to a preset correspondence between user attribute information and a basic attribute number; a second determining module 202, configured to determine, based on the first target attribute number, a target attribute set corresponding to the first target attribute number, where the target attribute set includes at least part of user attribute information in which a degree of association in the target attribute information is greater than a threshold; the processing module 203 is configured to perform normalization processing on at least part of the user attribute information to obtain feature attribute information corresponding to at least part of the user attribute information; the input module 204 is configured to input the feature attribute information to the trained prediction model, and obtain a product prediction result corresponding to the target user; and the recommending module 205 is configured to recommend a product corresponding to the product predicting result to the target user according to the product predicting result.
In a possible implementation manner, the first determining module 201 is further configured to extend a preset step size to a first direction and a second direction of the attribute number set with the first target attribute number as a starting point, so as to obtain a second target attribute number related to the first target attribute number; and determining that the user attribute information corresponding to the second target attribute number in the target attribute information belongs to the target attribute set.
In a possible implementation manner, the second determining module 202 is further configured to number the user attribute information as sample data, to obtain a first basic attribute number corresponding to the user attribute information; determining an attribute set corresponding to the first basic attribute number based on the first basic attribute number, wherein the attribute set comprises at least part of user attribute information with the association degree larger than a threshold value in the user attribute information, and carrying out normalization processing on at least part of the user attribute information in the attribute set to obtain user characteristic attribute information corresponding to at least part of the user attribute information in the attribute set; sampling data from the user characteristic attribute information based on a target mode to obtain a plurality of sample data sets, wherein each sample data set comprises at least one user characteristic attribute information; constructing a decision tree corresponding to a sample data set, and combining the decision trees to obtain a random forest flow prediction model to be trained; and training the random forest flow prediction model by utilizing the plurality of sample data sets to obtain a trained random forest flow prediction model.
In a possible implementation manner, the second determining module 202 is further configured to extend a preset step size to a first direction and a second direction of the attribute number set with the first basic attribute number as a starting point, obtain a second basic attribute number related to the first basic attribute number, and determine that user attribute information corresponding to the second basic attribute number belongs to the attribute set.
In a possible implementation manner, the input module 204 is further configured to input the feature attribute information into the trained prediction model, obtain probability values of prediction results output by each decision tree in the prediction model, sort the probability values in a descending order, and determine that the prediction results ranked in the first N are product prediction results.
In a possible implementation manner, the second determining module 202 is further configured to determine that the top N prediction results are product prediction results when the prediction results output by the decision trees in the trained prediction model are consistent with the prediction results output by the predetermined number of decision trees in the prediction model.
The product recommending device provided by the embodiment of the application can realize each process in the embodiment corresponding to the product recommending method, and has the same or similar beneficial effects, and in order to avoid repetition, the description is omitted here.
It should be noted that, the product recommending device provided in the embodiment of the present application and the product recommending method provided in the embodiment of the present application are based on the same application concept, and the product recommending device and the product recommending method are based on the same application concept, so that the specific implementation of the embodiment may refer to the implementation of the foregoing product recommending method, and have the same or similar beneficial effects, and the repetition is omitted.
According to the product recommendation method provided by the embodiment, based on the same technical concept, the embodiment of the application further provides an electronic device, which is used for executing the product recommendation method, and fig. 3 is a schematic structural diagram of an electronic device for implementing the embodiments of the application, as shown in fig. 3. The electronic device may be configured or configured differently, may include one or more processors 301 and memory 302, and may have one or more applications or data stored in memory 302. Wherein the memory 302 may be transient storage or persistent storage. The application programs stored in memory 302 may include one or more modules (not shown), each of which may include a series of computer-executable instructions for use in an electronic device.
Still further, the processor 301 may be arranged to communicate with the memory 302 and execute a series of computer executable instructions in the memory 302 on an electronic device. The electronic device may also include one or more power supplies 303, one or more wired or wireless network interfaces 304, one or more input/output interfaces 305, and one or more keyboards 306.
In this embodiment, the electronic device includes a processor, a communication interface, a memory, and a communication bus; the processor, the communication interface and the memory complete communication with each other through a bus; a memory for storing a computer program; the processor is configured to execute the program stored in the memory, implement each step in the above method embodiments, and have the beneficial effects of the above method embodiments, and in order to avoid repetition, the embodiments of the present application are not described herein again.
The embodiment also provides a computer readable storage medium, on which a computer program is stored, which when executed by a processor, implements each step in the above method embodiments, and has the beneficial effects of the above method embodiments, and in order to avoid repetition, the embodiments of the application are not described herein again.
It will be apparent to those skilled in the art that embodiments of the present application may be provided as a method, apparatus, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
In one typical configuration, the electronic device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
The memory may include volatile memory in a computer-readable medium, random Access Memory (RAM) and/or nonvolatile memory, etc., such as Read Only Memory (ROM) or flash memory (flashRAM). Memory is an example of computer-readable media.
Computer readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. Computer-readable media, as defined herein, does not include transitory computer-readable media (transshipment) such as modulated data signals and carrier waves.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or apparatus that comprises an element.
It will be apparent to those skilled in the art that embodiments of the present application may be provided as a method, apparatus, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The foregoing is merely exemplary of the present application and is not intended to limit the present application. Various modifications and changes may be made to the present application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc. which are within the spirit and principles of the present application are intended to be included within the scope of the claims of the present application.
Claims (10)
1. A product recommendation method, characterized in that the product recommendation method comprises:
under the condition that target attribute information of a target user is obtained, determining a first target attribute number of the target attribute information according to a corresponding relation between preset user attribute information and a basic attribute number;
determining a target attribute set corresponding to the first target attribute number based on the first target attribute number, wherein the target attribute set comprises at least part of user attribute information with association degree larger than a threshold value in the target attribute information;
normalizing the at least part of user attribute information to obtain feature attribute information corresponding to the at least part of user attribute information;
inputting the characteristic attribute information into a trained prediction model to obtain a product prediction result corresponding to the target user;
and recommending the product corresponding to the product prediction result to the target user according to the product prediction result.
2. The product recommendation method according to claim 1, wherein the determining a target attribute set corresponding to the first target attribute number based on the first target attribute number comprises:
respectively expanding a preset step length to a first direction and a second direction of an attribute number set by taking the first target attribute number as a starting point to obtain a second target attribute number related to the first target attribute number;
and determining that user attribute information corresponding to the second target attribute number in the target attribute information belongs to the target attribute set.
3. The product recommendation method according to claim 1, wherein the prediction model is a random forest flow prediction model, and before the feature attribute information is input into the trained prediction model to obtain a product prediction result corresponding to the target user, the method further comprises:
numbering the user attribute information serving as sample data to obtain a first basic attribute number corresponding to the user attribute information;
determining an attribute set corresponding to the first basic attribute number based on the first basic attribute number, wherein the attribute set comprises at least part of user attribute information with the association degree larger than a threshold value in the user attribute information, and carrying out normalization processing on at least part of user attribute information in the attribute set to obtain user characteristic attribute information corresponding to at least part of user attribute information in the attribute set;
data sampling is carried out from the user characteristic attribute information based on a target mode, so that a plurality of sample data sets are obtained, and each sample data set comprises at least one user characteristic attribute information;
constructing a decision tree corresponding to one sample data set, and combining the decision trees to obtain a random forest flow prediction model to be trained;
and training the random forest flow prediction model by utilizing each of the plurality of sample data sets to obtain the trained random forest flow prediction model.
4. The product recommendation method according to claim 3, wherein said determining an attribute set corresponding to the first basic attribute number based on the first basic attribute number comprises:
and respectively expanding a preset step length to a first direction and a second direction of an attribute number set by taking the first basic attribute number as a starting point to obtain a second basic attribute number related to the first basic attribute number, and determining that the user attribute information corresponding to the second basic attribute number belongs to the attribute set.
5. The product recommendation method according to claim 1, wherein the inputting the characteristic attribute information into the trained prediction model to obtain the product prediction result corresponding to the target user comprises:
and inputting the characteristic attribute information into a trained prediction model, acquiring probability values of prediction results output by each decision tree in the prediction model, ordering the probability values in a descending order, and determining that the prediction results ordered in the first N are the product prediction results.
6. The product recommendation method of claim 5, wherein said determining the top N ranked predictors as the product predictor comprises:
and under the condition that the predicted results output by all the decision trees in the trained predicted model are consistent with the predicted results output by the preset number of decision trees in the predicted model, determining the predicted results sequenced in the first N as the product predicted results.
7. A product recommendation device, characterized in that the product recommendation device comprises:
the first determining module is used for determining a first target attribute number of the target attribute information according to the corresponding relation between the preset user attribute information and the basic attribute number under the condition that the target attribute information of the target user is acquired;
a second determining module, configured to determine, based on the first target attribute number, a target attribute set corresponding to the first target attribute number, where the target attribute set includes at least part of user attribute information in which a degree of association in the target attribute information is greater than a threshold value;
the processing module is used for carrying out normalization processing on the at least part of user attribute information to obtain feature attribute information corresponding to the at least part of user attribute information;
the input module is used for inputting the characteristic attribute information into the trained prediction model to obtain a product prediction result corresponding to the target user;
and the recommending module is used for recommending the product corresponding to the product predicting result to the target user according to the product predicting result.
8. The product recommendation device of claim 7, wherein the second determining module is further configured to extend a preset step size in a first direction and a second direction of the attribute number set with the first target attribute number as a starting point, to obtain a second target attribute number related to the first target attribute number; and determining that user attribute information corresponding to the second target attribute number in the target attribute information belongs to the target attribute set.
9. An electronic device comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory perform communication with each other via the communication bus, the memory is configured to store a computer program, and the processor is configured to execute the program stored on the memory to implement the product recommendation method steps according to any one of claims 1-6.
10. A computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the product recommendation method steps of any of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210636020.2A CN117252656A (en) | 2022-06-07 | 2022-06-07 | Product recommendation method and device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210636020.2A CN117252656A (en) | 2022-06-07 | 2022-06-07 | Product recommendation method and device and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117252656A true CN117252656A (en) | 2023-12-19 |
Family
ID=89128112
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210636020.2A Pending CN117252656A (en) | 2022-06-07 | 2022-06-07 | Product recommendation method and device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117252656A (en) |
-
2022
- 2022-06-07 CN CN202210636020.2A patent/CN117252656A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108920654B (en) | Question and answer text semantic matching method and device | |
| CN109582793B (en) | Model training method, customer service system, data labeling system and readable storage medium | |
| CN113297396B (en) | Method, device and equipment for updating model parameters based on federal learning | |
| CN109543190A (en) | A kind of intension recognizing method, device, equipment and storage medium | |
| CN116596095B (en) | Training method and device of carbon emission prediction model based on machine learning | |
| CN111078876A (en) | Short text classification method and system based on multi-model integration | |
| AU2022201193A1 (en) | System and method for designing artificial intelligence (ai) based hierarchical multi-conversation system | |
| CN111737464B (en) | Text classification method and device and electronic equipment | |
| CN113656699B (en) | User feature vector determining method, related equipment and medium | |
| US10984781B2 (en) | Identifying representative conversations using a state model | |
| CN109597982B (en) | Abstract text recognition method and device | |
| CN107392311A (en) | The method and apparatus of sequence cutting | |
| CN114332550A (en) | Model training method, system, storage medium and terminal equipment | |
| CN108595395B (en) | Nickname generation method, device and equipment | |
| CN114792256B (en) | Crowd expansion method and device based on model selection | |
| CN117252656A (en) | Product recommendation method and device and electronic equipment | |
| CN112463964B (en) | Text classification and model training method, device, equipment and storage medium | |
| CN117312549A (en) | Work order processing method and device and electronic equipment | |
| CN114093447A (en) | Data asset recommendation method and device, computer equipment and storage medium | |
| CN114513578A (en) | Outbound method, device, computer equipment and storage medium | |
| CN114359649A (en) | Image processing method, apparatus, device, storage medium, and program product | |
| CN113704466A (en) | Text multi-label classification method and device based on iterative network and electronic equipment | |
| JP5824429B2 (en) | Spam account score calculation apparatus, spam account score calculation method, and program | |
| CN110059219A (en) | A kind of video preference prediction technique, device, equipment and readable storage medium storing program for executing | |
| CN114817695B (en) | Information recommendation method and device, electronic equipment and storage medium |
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 |