CN109299994A - Recommended method, device, equipment and readable storage medium storing program for executing - Google Patents
Recommended method, device, equipment and readable storage medium storing program for executing Download PDFInfo
- Publication number
- CN109299994A CN109299994A CN201810843390.7A CN201810843390A CN109299994A CN 109299994 A CN109299994 A CN 109299994A CN 201810843390 A CN201810843390 A CN 201810843390A CN 109299994 A CN109299994 A CN 109299994A
- Authority
- CN
- China
- Prior art keywords
- user
- application scenarios
- preference
- data
- under
- 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.)
- Granted
Links
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
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Marketing (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Business, Economics & Management (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention provides a kind of user preference recommended method, device, equipment and readable storage medium storing program for executing, extract the user characteristic data under each application scenarios in the historical data；Acquire the feature vector of historical review data under each application scenarios；Similarity parameter is calculated for each application scenarios using described eigenvector；Using the user characteristic data and the similarity parameter, the preference of each application scenarios is generated；Recommended based on the preference under each application scenarios.Solve the problems, such as that the Demand perference that can not achieve under the general scene for user in the prior art is recommended.
Description
Technical field
The present embodiments relate to electronic technology field more particularly to a kind of user preference recommended method, device, equipment and
Readable storage medium storing program for executing.
Background technique
While social application and shopping application are universal, generate a large amount of user data, wherein according to user data into
Row user preference statistics is the method that various social applications and shopping application generally use, to extract user characteristics interest, in turn
It is embodied as the purpose that user provides personalized service.
In the prior art, preference of the user to businessman, example are mainly excavated by user characteristics and the feature of businessman's dimension
Such as, by the microblogging text data of processing user, the higher word of the frequency of occurrences in text is determined as microblogging candidate topics, as
The expression of user interest profile, and the microblogging with similar topic is recommended into user.
However, the prior art can not recommend for the Demand perference under the general scene of user.
Summary of the invention
The present invention provides a kind of user preference recommended method, to solve to can not achieve the general field for user in first technology
The problem of Demand perference under scape is recommended.
According to the first aspect of the invention, a kind of user preference recommended method is provided, which comprises
The user characteristic data under each application scenarios is extracted in the historical data；
Acquire the feature vector of historical review data under each application scenarios；
Similarity parameter is calculated for each application scenarios using described eigenvector；
Using the user characteristic data and the similarity parameter, the preference of each application scenarios is generated；
Recommended based on the preference under each application scenarios.
According to the second aspect of the invention, a kind of user preference recommendation apparatus is provided, described device includes:
User characteristic data extraction module, for extracting the user characteristic data under each application scenarios in the historical data；
Feature vector acquisition module is commented on, for acquiring the feature vector of historical review data under each application scenarios；
Similarity parameter acquisition module, for calculating similarity parameter for each application scenarios using described eigenvector；
Application scenarios preference generation module is generated for using the user characteristic data and the similarity parameter
The preference of each application scenarios；
Preference recommending module, for being recommended based on the preference under each application scenarios.
According to the third aspect of the invention we, a kind of equipment is provided, comprising:
Processor, memory and it is stored in the computer journey that can be run on the memory and on the processor
Sequence, which is characterized in that the processor realizes user preference recommended method as the aforementioned when executing described program.
According to the fourth aspect of the invention, provide a kind of readable storage medium storing program for executing, when the instruction in the storage medium by
When the processor of electronic equipment executes, so that electronic equipment is able to carry out user preference recommended method above-mentioned.
A kind of user preference recommended method, device, equipment and readable storage medium storing program for executing provided in an embodiment of the present invention, in history
The user characteristic data under each application scenarios is extracted in data；Acquire the feature vector of historical review data under each application scenarios；
Similarity parameter is calculated for each application scenarios using described eigenvector；Using the user characteristic data and the similarity
Parameter generates the preference of each application scenarios；Recommended based on the preference under each application scenarios.This motion passes through needle
The similarity parameter of preference and each application scenarios to a application scenarios realizes to user and recommends preference application scenarios
Purpose.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, below by institute in the description to the embodiment of the present invention
Attached drawing to be used is needed to be briefly described, it should be apparent that, the accompanying drawings in the following description is only some implementations of the invention
Example, for those of ordinary skill in the art, without any creative labor, can also be according to these attached drawings
Obtain other attached drawings.
Fig. 1 is a kind of step flow chart of user preference recommended method provided in an embodiment of the present invention；
Fig. 2 is a kind of step flow chart of user preference recommended method provided in an embodiment of the present invention；
Fig. 2A is that user preference recommended method provided in an embodiment of the present invention realizes structural schematic diagram；
Fig. 2 B is schematic diagram of decaying between the user's history Behavior preference added-time provided in an embodiment of the present invention；
Fig. 2 C is user's scene preference sparse matrix schematic diagram provided in an embodiment of the present invention；
Fig. 2 D is individual scene drainage effect Statistical Comparison schematic diagram provided in an embodiment of the present invention；
Fig. 3 is a kind of structure chart of user preference recommendation apparatus provided in an embodiment of the present invention；
Fig. 4 is a kind of structure chart of user preference recommendation apparatus provided in an embodiment of the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are some of the embodiments of the present invention, instead of all the embodiments.Based on this hair
Embodiment in bright, every other implementation obtained by those of ordinary skill in the art without making creative efforts
Example, shall fall within the protection scope of the present invention.
The term being related in the embodiment of the present invention is introduced first below:
LDA (hidden Di Li Cray distributed model) is a kind of document subject matter generation model, includes word, theme and three layers of document
Structure extracts theme using statistic sampling method mainly for the treatment of extensive text cluster.
Word2vec, be it is a kind of words will be switched to the vector field homoemorphism type that computer is understood that, mainly for the treatment of short
The vector of text indicates problem
Collaborative Recommendation is a kind of method for recommending it similar purchase article by finding similar users, in this motion
The matrix of user and theme is filled using ALS (alternately least square) method, solves the problems, such as the sparsity of user's matrix.
Embodiment one
Referring to Fig.1, it illustrates a kind of step flow charts of user preference recommended method, the specific steps of which are as follows:
Step 101, the user characteristic data under each application scenarios is extracted in the historical data；
In the embodiment of the present invention, in using each user's history data recorded in background data base, feature extraction is used
Algorithm, such as LDA carry out feature extraction for the comment data that each application scenarios are delivered to user, and utilize word2vec pairs
The historical search record for each application scenarios of user carries out feature extraction, further according to the gender of user, age, occupation, wedding
The sparse preference of user, and generation pair are filled and made up to the essential informations such as relation by marriage state, user's history user behaviors log and Collaborative Recommendation
Answer the user characteristic data of each application scenarios.
Step 102, the feature vector of historical review data under each application scenarios is acquired.
In the embodiment of the present invention, application scenarios are often referred to businessman's scene, are acquiring the historical review number in each application scenarios
According to rear, a unified text is generated, recycles text feature algorithm (doc2vec) to obtain each application scenarios and unifies comment text
Feature vector.
Step 103, similarity parameter is calculated for each application scenarios using described eigenvector；
In the embodiment of the present invention, using the comment text feature vector of each application scenarios of correspondence as the similar of each application scenarios
Degree, the mark of user's scene preference, i.e. the similarity parameter of application scenarios are converted to user's businessman's Behavior preference.
Step 104, using the user characteristic data and the similarity parameter, the preference of each application scenarios is generated.
In the embodiment of the present invention, the similarity obtained in the user characteristic data and step 103 that step 101 obtains
Parameter, unbalanced input disaggregated model, such as support vector machines, the value of output are then the preference that user is directed to each application scenarios
Degree.
Step 105, recommended based on the preference under each application scenarios.
In the embodiment of the present invention, according to the preference of the application scenarios of the user, recommend different application scenarios to user.
Wherein, recommend application scenarios that can carry out cis-position displaying according to the preference power of user, or only revealed preference degree is strongest
Application scenarios, the present invention are without restriction to this.
In conclusion a kind of user preference recommended method provided in an embodiment of the present invention, extracts respectively answer in the historical data
With the user characteristic data under scene；Acquire the feature vector of historical review data under each application scenarios；Using the feature to
Metering pin calculates similarity parameter to each application scenarios；Using the user characteristic data and the similarity parameter, generation is respectively answered
Recommended with the preference of scene based on the preference under each application scenarios.This motion is directed to an applied field by user
The similarity parameter of the preference of scape and each application scenarios realizes finer to user's recommendation preference application scenarios
Purpose.
Embodiment two
Referring to Fig. 2, it illustrates a kind of step flow charts of user preference recommended method, the specific steps of which are as follows:
Step 201, by being carried out at LDA to the user in historical data in the historical review data under each application scenarios
Reason, extracts the comment theme distribution feature vector of user.
In the embodiment of the present invention, as shown in Figure 2 A, first part describes the process of user data preparation, wherein 1
It is the historical review data that each application scenarios are directed to by user in historical data that backstage will be applied to store, using LDA algorithm,
Carry out feature extraction.
Wherein, LDA (Latent Dirichlet Allocation) is that a kind of document subject matter generates model, also referred to as one
A three layers of bayesian probability model includes word, theme and document three-decker.So-called generation model, that is, if it is considered to one
Each word of piece article is by " with some theme of certain probability selection, and with certain probability selection from this theme
Such a process of some word " obtains.Document obeys multinomial distribution to theme, and theme to word obeys multinomial distribution.LDA
It is a kind of non-supervisory machine learning techniques, can be used to identify extensive document sets (document collection) or corpus
The subject information hidden in library (corpus).The method that it uses bag of words (bag of words), this method is by each piece
Document is considered as a word frequency vector, so that text information is converted the digital information for ease of modeling.Each documents representative
Probability distribution that some themes are constituted, and each theme represents the probability point that many words are constituted
Cloth.
Specifically, when user for each application scenarios historical review data after LDA is handled, each user can be obtained
For the comment word frequency vector of each application scenarios, which maps out user to the inclined of each application scenarios to a certain extent
It is good.
Step 202, word2vec processing is carried out by the historical search data to the user in historical data, obtains user
Search for term vector.
In the embodiment of the present invention, in first part as shown in Figure 2 A, wherein 2 be by obtaining user in historical data
Historical search data, using word2vec algorithm, obtains search term using all historical search datas of user as process object
Vector.
Wherein, word2vec is also word embeddings, and Chinese name " term vector ", effect is exactly will be in natural language
Words switch to the dense vector (Dense Vector) that computer is understood that.Even if corpus is not sufficient enough, obtained word
Vector matrix still will not be sparse, and can show the relationship between each vector.
Specifically, when the historical search word of user is as corpus, the search term handled using word2vec
Vector, using user comment training word2vec model, mainly for the treatment of the correlation of short text；To user's history search term
(query) it is segmented, term vector is found by word2vec respectively, and aggregate into total user's query vector, mathematical notation
It is as follows:
Query 1: vector M (a1,a2,a3,…..aN)
Query 2: vector N (b1,b2,b3,…bN)
Query 3: vector Q (c1,c2,c3,…cN)
Always
This method can reinforce the influence of different query, such as " peppery chafing dish " on different dimensions, respectively in some table
Show that score value is very high in peppery dimension, user can be reinforced for peppery preference with above method adduction.
Step 203, in conjunction with user base information, the user in historical data comment theme distribution feature vector,
The user searches for term vector, generates the user characteristic data that user is directed to each application scenarios.
In the embodiment of the present invention, user base information in as shown in Figure 2 A 3, such as gender, age, occupation, marriage
The information such as state, state in love, local, the comment theme distribution feature vector of the user in conjunction with obtained in step 201-202, institute
It states user and searches for term vector, above-mentioned data be integrated into a whole user for each application scenarios feature vector, integration
Data, which can be, to be stored in the matrix form and is further processed, and certainly, for different processing method or model, integrates number
According to storage mode may be different, the embodiments of the present invention are not limited thereto.
Preferably, further includes:
Step S1 obtains other use under each application scenarios by the marked high trusted users data in historical data
The reinforcing Collaborative Recommendation data at family.
In the embodiment of the present invention, as shown in 5 in Fig. 2A first part, using collaborative recommendation method filling user to theme
Preference sparse matrix (as shown in Figure 2 C), wherein U is user, I theme, for solve the problems, such as user to the sparsity of theme, this
In the ALS method that uses it is similar with other general collaborative recommendation methods, cardinal principle be recommended by multiple similar users it is similar
As preference.
Preferably, step S1 includes sub-step A1-A5；
Sub-step A1 obtains high trusted users data marked under each application scenarios in historical data.
It to be high trusted users by some user's marks that each application scenarios are directed in the embodiment of the present invention, in historical data,
For example, the user for often generating user behavior under the application scenarios, and enjoying a good reputation.
Sub-step A2 obtains the high trusted users to the preference of each application scenarios by high trusted users data
Degree.
In the embodiment of the present invention, by the processing to high trusted users data, such as to all comment numbers of high trusted users
According to processing, preference of the available high trusted users to each application scenarios.
Sub-step A3 extracts the historical behavior log of high trusted users in the high trusted users data.
In the embodiment of the present invention, the historical behavior log of high trusted users is extracted, and further obtains each high trusted users
Behavior vector.
Sub-step A4 generates institute according to the historical behavior log of the high trusted users using alternately leastsquares algorithm
State behavior vector characteristics of the high trusted users under each application scenarios.
In the embodiment of the present invention, using alternately leastsquares algorithm ALS, obtain the feature of the behavior of each high trusted users to
Amount, such as high trusted users are A, B, C, then their vectors of their behavior are vecA, vecB, vecC。
Sub-step A5 obtains the other users pair in historical data according to the preference and the behavior vector characteristics
The preference of each application scenarios.
In the embodiment of the present invention, strengthening behavior cooperates with concrete practice method as follows:
High trusted users A, B, the C of certain known scene, corresponding scene preference are prefA, prefB,prefc,
Their ALS behavior vector is vecA, vecB, vecC, then other users (such as user N) calculate the preference of scene are as follows:
Thus, it is possible to obtain in historical data other users for the preference of each application scenarios.
Preferably, further includes:
Step S2 generates user's needle in conjunction with the user base information and the reinforcing Collaborative Recommendation data in historical data
To the user characteristic data of each application scenarios.
In the embodiment of the present invention, user base information in as shown in Figure 2 A 3, such as gender, age, occupation, marriage
The information such as state, state in love, local, the comment theme distribution of the user in conjunction with obtained in step 201-202 and step S1
Feature vector, the user search for term vector, the user's history time of the act preference profiles and the reinforcing Collaborative Recommendation
The user that above-mentioned data be integrated into an entirety, is directed to the characteristic of each application scenarios by data, and integral data can be with
It is to store and be further processed in the matrix form, certainly, for different processing method or model, the storage of integral data
Mode may be different, and the embodiments of the present invention are not limited thereto.
Step 204, user's active time under each application scenarios is obtained.
In the embodiment of the present invention, according to the user's history user behaviors log of application backstage record, user can be extracted and be directed to
The active time section of each application scenarios, for example, user A is between 8. -10 points of every morning, in the stop of businessman's scenario B
Time longest.Wherein it is possible to carry out the extraction of application time by different statisticals, can also be obtained in conjunction with a variety of statisticals
Family active time is taken, the embodiments of the present invention are not limited thereto.
Preferably, step 204, further includes: sub-step B1- sub-step B2；
Sub-step B1 obtains the user's history user behaviors log in the historical data.
In the embodiment of the present invention, second part as shown in Figure 2 A is extracted in the historical data of application backstage storage and is used
Family historical behavior log, the user behavior being usually stored in the log in the line duration and line duration of user.
Sub-step B2, according to the user's history user behaviors log, obtain user's average active period under each application scenarios,
The active time of user's active time section and user under each application scenarios.
In the embodiment of the present invention, wherein user's average active period in user's history user behaviors log, user can be extracted
The active time of active time section and user under each application scenarios, as the major parameter for determining user activity, certainly,
In practical application, user activity judges the parameter for being not limited to foregoing description, and the embodiments of the present invention are not limited thereto.
Preferably, further includes:
The historical behavior log binding time weight of user in historical data is obtained user's history behavior by step V1
Time preference's feature；
In the embodiment of the present invention, in first part as shown in Figure 2 A, wherein 4 be using time decay factor as user
The weight of historical behavior feature, analog subscriber historical behavior is as time goes by the influence of user preference.
As shown in Figure 2 B, fusion user's history (in T days) is clicked, and is placed an order, and is bought, and collection is commented on behavior, and weighed
It sums it up again, calculation formula is as follows:
Wherein tiFor current date, thistoryFor the date that user behavior occurs, ti-thistoryOccur for user behavior
Length of the date apart from current date, the time is longer, and impact factor is smaller, for analog subscriber historical behavior pushing away with the time
Moving, which influences user preference, becomes smaller.
Step V2 is generated in conjunction with the user base information in historical data, the user's history time of the act preference profiles
User is directed to the user characteristic data of each application scenarios.
In the embodiment of the present invention, user base information in as shown in Figure 2 A 3, such as gender, age, occupation, marriage
Information, the user's history time of the act preference profiles in conjunction with obtained in step V1 such as state, state in love, local are integrated into
One whole user is directed to each application scenarios feature vector, and integral data, which can be, to be stored in the matrix form and carry out at next step
Reason, certainly, for different processing method or model, the storage mode of integral data may be different, and the embodiment of the present invention is to this
It is without restriction.
Step 205, the feature vector of historical review data under each application scenarios is acquired.
This step is identical as step 102, and this will not be detailed here.
Step 206, similarity parameter is calculated for each application scenarios using described eigenvector.
This step is identical as step 103, and this will not be detailed here.
Step 207, by the user characteristic data and the similarity parameter, using default Nonlinear Classification model and
Sort algorithm generates the preference of each application scenarios.
In the embodiment of the present invention, in Part III as shown in Figure 2 A, prediction model sorts to user preference scene, this Shen
Please used in be algorithm of support vector machine, in Fig. 2A first part generate user characteristic data and Part III generate
User's scene similarity parameter inputs supporting vector machine model, obtains the ordering of optimization preference that user is directed to each application scenarios, i.e. preference
Degree.
Wherein, support vector machines (Support-Vector Machines) is mainly used for classification and regression analysis.Given one
Group training sample, each label is two classes, and a SVM training algorithm establishes a model, and distributing new example is one
Class or other classes become non-probability binary linearity classification.The example of one SVM model, point such as in space, mapping,
So that it is the expression of division as wide as possible that the example of the different classification, which is by an apparent gap,.New embodiment is then reflected
It is mapped in identical space, and predicts to fall in the clearance side based on them and belong to a classification.It can most importantly incite somebody to action
Nonlinear Classification is effectively performed in implicit be mapped in high-dimensional feature space of input data.
Specifically, available user is directed to the phase of each application scenarios when input user characteristic data and similarity parameter
Classify like degree, recycle sort algorithm, application scenarios are ranked up according to similarity height, finally obtain user for respectively answering
With the preference of scene.
The user's active time being preferably based under each application scenarios and corresponding preference, are recommended
In the embodiment of the present invention, for user's active time of each application scenarios obtained in step 204, when detecting use
, then according to the preference of the application scenarios of the user, recommend different applied fields to user when corresponding active time is online in family
Scape.Wherein, recommend application scenarios that can carry out cis-position displaying according to the preference power of user, or only revealed preference degree is most strong
Application scenarios, the present invention is without restriction to this.
Step 208, in user's average active period, user's active time section and user under application scenarios when enlivening
In, at least one application scenarios is determined according to user search term currently entered.
In the embodiment of the present invention, after obtaining user activity, in user's active time section, i.e., user's average active is all
Phase, user's active time section and user are in the active time under application scenarios, when user is online, and input search key
When, according to the corresponding application scenarios of the search term of user, determine at least one application scenarios.For example, user's input " peppery, fire
Pot ", then be assured that application scenarios belong to " cuisines classification ", wherein businessman's scene is related to being confirmed as pair for chafing dish
The application scenarios answered.
Step 209, according to the preference of the application scenarios, at least one described application scenarios are carried out to the user
It shows.
In the embodiment of the present invention, the application scenarios preference according to obtained in step 208 sequence, by each application scenarios according to
Sequence shows user.Multiple application scenarios can be shown according to user setting using showing in interface, it can also be with revealed preference
Highest application scenarios are spent, it is without restriction to this embodiment of the present invention.
Specifically, user agent drainage realizes apparent growth, as shown in Figure 2 D by carrying out preference recommendation to user
, the customer flow growth guidance for each personalized channels significantly improves.
In conclusion a kind of user preference recommended method provided in an embodiment of the present invention, by the use in historical data
Historical review data of the family under each application scenarios carry out LDA processing, extract the comment theme distribution feature vector of user；Pass through
Word2vec processing is carried out to the historical search data of the user in historical data, user is obtained and searches for term vector；By history number
The historical behavior log binding time weight of user in obtains user's history time of the act preference profiles；Pass through history number
Marked high trusted users data in, obtain the reinforcing Collaborative Recommendation data of other users under each application scenarios；In conjunction with
User base information and the reinforcing Collaborative Recommendation data in historical data, the user for generating user for each application scenarios are special
Levy data；Acquire the feature vector of historical review data under each application scenarios；Each application scenarios are directed to using described eigenvector
Calculate similarity parameter；Using the user characteristic data and the similarity parameter, the preference of each application scenarios is generated；It obtains
Take user's active time under each application scenarios；Recommended based on the preference under each application scenarios.This motion passes through
User activity and user are directed to a preference for application scenarios and the similarity parameter of each application scenarios, realize more
The fine active time for user recommends the purpose of preference application scenarios to user.
Embodiment three
It is specific as follows it illustrates a kind of structural block diagram of user preference recommendation apparatus referring to Fig. 3:
User characteristic data extraction module 301, for extracting the user characteristics number under each application scenarios in the historical data
According to；
Feature vector acquisition module 302 is commented on, for acquiring the feature vector of historical review data under each application scenarios；
Similarity parameter acquisition module 303, for calculating similarity ginseng for each application scenarios using described eigenvector
Number；
Application scenarios preference generation module 304, it is raw for using the user characteristic data and the similarity parameter
At the preference of each application scenarios；
Preference recommending module 305, for being recommended based on the preference under each application scenarios.
Referring to Fig. 4, it illustrates the structural block diagram of another user preference recommendation apparatus based on Fig. 3 embodiment, tools
Body is as follows:
User characteristic data extraction module 301, for extracting the user characteristics number under each application scenarios in the historical data
According to；
Preferably, the user characteristic data extraction module 301.It specifically includes:
Comment on theme distribution characteristic vector pickup submodule 3011, for by the user in historical data in each application
Historical review data under scene carry out LDA processing, extract the comment theme distribution feature vector of user；
User searches for term vector acquisition submodule 3012, for passing through the historical search data to the user in historical data
Word2vec processing is carried out, user is obtained and searches for term vector；
First user characteristics generate submodule 3013, in conjunction with the user base information in historical data, the user
Comment theme distribution feature vector, the user search for term vector, generate user be directed to each application scenarios user characteristics number
According to；
Preferably, the user characteristic data extraction module 301, further includes:
Strengthen Collaborative Recommendation data acquisition submodule, for passing through the marked high trusted users number in historical data
According to obtaining the reinforcing Collaborative Recommendation data of other users under each application scenarios；
Second user feature generates submodule, for cooperateing in conjunction with the user base information in historical data with the reinforcing
Recommending data generates the user characteristic data that user is directed to each application scenarios
Preferably, the reinforcing Collaborative Recommendation data acquisition submodule, comprising:
High trusted users data capture unit, can credit for obtaining height marked under each application scenarios in historical data
User data；
Application scenarios preference acquiring unit, for obtaining the high trusted users to institute by high trusted users data
State the preference of each application scenarios；
Historical behavior log acquisition unit, for extracting the history row of high trusted users in the high trusted users data
For log；
Behavior vector characteristics generation unit, for using alternately leastsquares algorithm, according to going through for the high trusted users
History user behaviors log generates behavior vector characteristics of the high trusted users under each application scenarios；
Preference acquiring unit, for obtaining in historical data according to the preference and the behavior vector characteristics
Preference of the other users to each application scenarios.
Preferably, the user characteristic data extraction module 301, further includes:
Time preference's feature acquisition submodule, for weighing the historical behavior log binding time of the user in historical data
Weight obtains user's history time of the act preference profiles；
Third user characteristics generate submodule, in conjunction with user base information, the user's history in historical data
Time of the act preference profiles generate the user characteristic data that user is directed to each application scenarios.
Feature vector acquisition module 302 is commented on, for acquiring the feature vector of historical review data under each application scenarios；
Similarity parameter acquisition module 303, for calculating similarity ginseng for each application scenarios using described eigenvector
Number；
Application scenarios preference generation module 304, it is raw for using the user characteristic data and the similarity parameter
At the preference of each application scenarios；
Preferably, the application scenarios preference generation module 304, comprising:
Application scenarios preference generates submodule 3041, for being joined by the user characteristic data and the similarity
Number generates the preference of each application scenarios using default Nonlinear Classification model and sort algorithm.
User's active time obtains module 305, for obtaining user's active time under each application scenarios；
Preferably, user's active time obtains module 305, comprising:
User's history user behaviors log acquisition submodule, for obtaining the user's history user behaviors log in the historical data；
Active time acquisition submodule, for obtaining the use under each application scenarios according to the user's history user behaviors log
The active time of family average active period, user's active time section and user under each application scenarios.
Preference recommending module 306, for being recommended based on the preference under each application scenarios.
Preferably, the preference recommending module 306, comprising:
Application scenarios determine submodule 3061, in user's average active period, user's active time section and user
In the active time under application scenarios, at least one application scenarios is determined according to user search term currently entered；
Recommend to show submodule 3062, it, will at least one described applied field for the preference according to the application scenarios
Scape is shown to the user.
The embodiment of the present invention also provides a kind of equipment, comprising: processor, memory and is stored on the memory simultaneously
The computer program that can be run on the processor, which is characterized in that the processor is realized as above when executing described program
User preference recommended method described in the one or more stated.
The embodiment of the present invention also provides a kind of readable storage medium storing program for executing, when the instruction in the storage medium is by electronic equipment
When processor executes, so that electronic equipment is able to carry out user preference recommended method as mentioned.
Preferably, further includes:
Active time recommending module, for based under each application scenarios user's active time and corresponding preference
Degree, is recommended.
In conclusion a kind of user preference recommended method provided in an embodiment of the present invention, is extracted by user characteristic data
Module, for extracting the user characteristic data under each application scenarios in the historical data；Secondly, passing through comment feature vector acquisition
Module, for acquiring the feature vector of historical review data under each application scenarios；Again by similarity parameter acquisition module, use
Similarity parameter is calculated in being directed to each application scenarios using described eigenvector；Later, it by similarity parameter acquisition module, uses
In using the user characteristic data and the similarity parameter, the preference of each application scenarios is generated；And user is when enlivening
Between obtain module, for obtaining user's active time under each application scenarios；Finally, by preference recommending module, for being based on
User's active time and corresponding preference under each application scenarios carry out user preference and recommend to be based on each applied field
Preference under scape is recommended.This motion is directed to the preference of an application scenarios by user activity and user, and each
The similarity parameter of application scenarios realizes the finer active time for user, recommends preference applied field to user
The purpose of scape.It has the advantages that
First, being weighted the mode of polymerization to user's businessman's preference, by businessman's dimension preference to convert scene dimension inclined
It is good, it is more efficient that user's scene preference has accurately been determined.
Second, application of the LDA theme vector in the excavation of user version preference maximizes user in conjunction with user query
The utilization of the text data for the preference that directly shows off.
Thirdly, sparse behavior to businessman of ALS extension and filling user, and pass through strengthening behavior collaboration extended scene preference
User.
Described herein recommend, including recommending search term, businessman, commodity etc. under associated scenario to user.
By more accurately recommending to user, user can be triggered and more click behavior, and the more order conversions of output in turn.
For device embodiment, since it is basically similar to the method embodiment, related so being described relatively simple
Place illustrates referring to the part of embodiment of the method.
Algorithm and display are not inherently related to any particular computer, virtual system, or other device provided herein.
Various general-purpose systems can also be used together with teachings based herein.As described above, it constructs required by this kind of system
Structure be obvious.In addition, the present invention is also not directed to any particular programming language.It should be understood that can use various
Programming language realizes summary of the invention described herein, and the description done above to language-specific is to disclose this hair
Bright preferred forms.
In the instructions provided here, numerous specific details are set forth.It is to be appreciated, however, that implementation of the invention
Example can be practiced without these specific details.In some instances, well known method, structure is not been shown in detail
And technology, so as not to obscure the understanding of this specification.
Similarly, it should be understood that in order to simplify the disclosure and help to understand one or more of the various inventive aspects,
Above in the description of exemplary embodiment of the present invention, each feature of the invention is grouped together into single implementation sometimes
In example, figure or descriptions thereof.However, the disclosed method should not be interpreted as reflecting the following intention: i.e. required to protect
Shield the present invention claims features more more than feature expressly recited in each claim.More precisely, as following
Claims reflect as, inventive aspect is all features less than single embodiment disclosed above.Therefore,
Thus the claims for following specific embodiment are expressly incorporated in the specific embodiment, wherein each claim itself
All as a separate embodiment of the present invention.
Those skilled in the art will understand that can be carried out adaptively to the module in the equipment in embodiment
Change and they are arranged in one or more devices different from this embodiment.It can be the module or list in embodiment
Member or component are combined into a module or unit or component, and furthermore they can be divided into multiple submodule or subelement or
Sub-component.Other than such feature and/or at least some of process or unit exclude each other, it can use any
Combination is to all features disclosed in this specification (including adjoint claim, abstract and attached drawing) and so disclosed
All process or units of what method or apparatus are combined.Unless expressly stated otherwise, this specification is (including adjoint power
Benefit require, abstract and attached drawing) disclosed in each feature can carry out generation with an alternative feature that provides the same, equivalent, or similar purpose
It replaces.
Various component embodiments of the invention can be implemented in hardware, or to run on one or more processors
Software module realize, or be implemented in a combination thereof.It will be understood by those of skill in the art that can be used in practice
Microprocessor or digital signal processor (DSP) realize one in payment information processing equipment according to an embodiment of the present invention
The some or all functions of a little or whole components.The present invention is also implemented as executing method as described herein
Some or all device or device programs (for example, computer program and computer program commodity data).It is such
It realizes that program of the invention can store on a computer-readable medium, or can have the shape of one or more signal
Formula.Such signal can be downloaded from an internet website to obtain, and perhaps be provided on the carrier signal or with any other shape
Formula provides.
It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and ability
Field technique personnel can be designed alternative embodiment without departing from the scope of the appended claims.In the claims,
Any reference symbol between parentheses should not be configured to limitations on claims.Word "comprising" does not exclude the presence of not
Element or step listed in the claims.Word "a" or "an" located in front of the element does not exclude the presence of multiple such
Element.The present invention can be by means of including the hardware of several different elements and being come by means of properly programmed computer real
It is existing.In the unit claims listing several devices, several in these devices can be through the same hardware branch
To embody.The use of word first, second, and third does not indicate any sequence.These words can be explained and be run after fame
Claim.
It is apparent to those skilled in the art that for convenience and simplicity of description, the system of foregoing description,
The specific work process of device and unit, can refer to corresponding processes in the foregoing method embodiment, and details are not described herein.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention
Made any modifications, equivalent replacements, and improvements etc., should all be included in the protection scope of the present invention within mind and principle.
The above description is merely a specific embodiment, but scope of protection of the present invention is not limited thereto, any
Those familiar with the art in the technical scope disclosed by the present invention, can easily think of the change or the replacement, and should all contain
Lid is within protection scope of the present invention.Therefore, protection scope of the present invention should be subject to the protection scope in claims.
Claims (12)
1. a kind of recommended method, which is characterized in that the described method includes:
The user characteristic data under each application scenarios is extracted in the historical data；
Acquire the feature vector of historical review data under each application scenarios；
Similarity parameter is calculated for each application scenarios using described eigenvector；
Using the user characteristic data and the similarity parameter, the preference of each application scenarios is generated；
Recommended based on the preference under each application scenarios.
2. the method according to claim 1, wherein the use extracted under each application scenarios in the historical data
The step of family characteristic, comprising:
By carrying out LDA processing to historical review data of the user in historical data under each application scenarios, extract user's
Comment on theme distribution feature vector；
Word2vec processing is carried out by the historical search data to the user in historical data, user is obtained and searches for term vector；
It is searched in conjunction with the comment theme distribution feature vector of user base information, the user in historical data, the user
Term vector generates the user characteristic data that user is directed to each application scenarios.
3. the method according to claim 1, wherein the use extracted under each application scenarios in the historical data
The step of family characteristic, comprising:
By the marked high trusted users data in historical data, the reinforcing collaboration of other users under each application scenarios is obtained
Recommending data；
In conjunction with the user base information and the reinforcing Collaborative Recommendation data in historical data, generates user and be directed to each application scenarios
User characteristic data.
4. the method according to claim 1, wherein further include: when obtaining the user under each application scenarios and enlivening
Between；And
Based under each application scenarios user's active time and corresponding preference, recommended.
5. according to the method described in claim 4, it is characterized in that, the use extracted under each application scenarios in the historical data
The step of family characteristic, comprising:
By the historical behavior log binding time weight of the user in historical data, it is special to obtain user's history time of the act preference
Sign；
In conjunction with the user base information in historical data, the user's history time of the act preference profiles, user is generated for each
The user characteristic data of application scenarios.
6. according to the method described in claim 3, it is characterized in that, described pass through high trusted users marked in historical data
Data, obtain under each application scenarios be directed to the user reinforcing Collaborative Recommendation data the step of, comprising:
Obtain high trusted users data marked under each application scenarios in historical data；
By high trusted users data, the high trusted users are obtained to the preference of each application scenarios；
The historical behavior log of high trusted users is extracted in the high trusted users data；
The high trusted users are generated according to the historical behavior log of the high trusted users using alternately leastsquares algorithm
Behavior vector characteristics under each application scenarios；
According to the preference and the behavior vector characteristics, the other users in historical data are obtained to each application scenarios
Preference.
7. the method according to claim 1, wherein described use the user characteristic data and the similarity
Parameter, the step of generating the preference of each application scenarios, comprising:
It is raw using default Nonlinear Classification model and sort algorithm by the user characteristic data and the similarity parameter
At the preference of each application scenarios.
8. according to the method described in claim 4, it is characterized in that, described obtain user's active time under each application scenarios
Step, comprising:
Obtain the user's history user behaviors log in the historical data；
According to the user's history user behaviors log, user's average active period under each application scenarios, user's active time are obtained
The active time of section and user under each application scenarios.
9. according to the method described in claim 8, it is characterized in that, the preference based under each application scenarios carries out
The step of recommendation, comprising:
In user's average active period, user's active time section and user in the active time under application scenarios, according to institute
It states user's search term currently entered and determines at least one application scenarios；
According to the preference of the application scenarios, at least one described application scenarios are shown to the user.
10. a kind of user preference recommendation apparatus, which is characterized in that described device includes:
User characteristic data extraction module, for extracting the user characteristic data under each application scenarios in the historical data；
Feature vector acquisition module is commented on, for acquiring the feature vector of historical review data under each application scenarios；
Similarity parameter acquisition module, for calculating similarity parameter for each application scenarios using described eigenvector；
Application scenarios preference generation module, for using the user characteristic data and the similarity parameter, generation is respectively answered
With the preference of scene；
Preference recommending module, for being recommended based on the preference under each application scenarios.
11. a kind of equipment characterized by comprising
Processor, memory and it is stored in the computer program that can be run on the memory and on the processor,
It is characterized in that, the processor realizes described in any item recommended methods such as claim 1-9 when executing described program.
12. a kind of readable storage medium storing program for executing, which is characterized in that when the instruction in the storage medium is held by the processor of electronic equipment
When row, so that electronic equipment is able to carry out described in any item recommended methods such as claim to a method 1-9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810843390.7A CN109299994B (en) | 2018-07-27 | 2018-07-27 | Recommendation method, device, equipment and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810843390.7A CN109299994B (en) | 2018-07-27 | 2018-07-27 | Recommendation method, device, equipment and readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109299994A true CN109299994A (en) | 2019-02-01 |
| CN109299994B CN109299994B (en) | 2021-09-24 |
Family
ID=65172668
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810843390.7A Active CN109299994B (en) | 2018-07-27 | 2018-07-27 | Recommendation method, device, equipment and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109299994B (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110866180A (en) * | 2019-10-12 | 2020-03-06 | 平安国际智慧城市科技股份有限公司 | Resource recommendation method, server and storage medium |
| CN111080339A (en) * | 2019-11-18 | 2020-04-28 | 口口相传(北京)网络技术有限公司 | Method and device for generating category preference data based on scene |
| CN111159393A (en) * | 2019-12-30 | 2020-05-15 | 电子科技大学 | Text generation method for abstracting abstract based on LDA and D2V |
| CN111368202A (en) * | 2020-03-06 | 2020-07-03 | 咪咕文化科技有限公司 | Search recommendation method and device, electronic equipment and storage medium |
| CN111553748A (en) * | 2020-05-09 | 2020-08-18 | 福州大学 | Android micro-service recommendation method and system based on user scene |
| CN111800569A (en) * | 2019-04-09 | 2020-10-20 | Oppo广东移动通信有限公司 | Photographing processing method and device, storage medium and electronic equipment |
| CN112749343A (en) * | 2021-01-22 | 2021-05-04 | 武汉蔚来能源有限公司 | Resource recommendation method and device and computer storage medium |
| WO2021082500A1 (en) * | 2019-10-31 | 2021-05-06 | 百果园技术(新加坡)有限公司 | Information recommendation method, apparatus, recommendation server and storage medium |
| CN113496432A (en) * | 2021-07-06 | 2021-10-12 | 北京爱笔科技有限公司 | Mining method, device and equipment of entity to be recommended and storage medium |
| CN114398486A (en) * | 2022-01-06 | 2022-04-26 | 北京博瑞彤芸科技股份有限公司 | Method and device for intelligently customizing customer acquisition publicity |
| CN114443967A (en) * | 2022-04-08 | 2022-05-06 | 北京并行科技股份有限公司 | Similar application recommendation method, computing device and storage medium |
| CN114528484A (en) * | 2022-01-26 | 2022-05-24 | 北京金堤科技有限公司 | Preference mining method and device, storage medium and electronic equipment |
| CN115470414A (en) * | 2022-11-03 | 2022-12-13 | 安徽商信政通信息技术股份有限公司 | United celebrity recommendation method and recommendation system |
| WO2023019517A1 (en) * | 2021-08-19 | 2023-02-23 | 阿波罗智联(北京)科技有限公司 | Instruction recommendation method and apparatus |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103106285A (en) * | 2013-03-04 | 2013-05-15 | 中国信息安全测评中心 | Recommendation algorithm based on information security professional social network platform |
| CN103324750A (en) * | 2013-07-04 | 2013-09-25 | 莫志鹏 | Method for personal screening of photo galleries on the basis of Bayesian network |
| CN105812937A (en) * | 2014-12-30 | 2016-07-27 | Tcl集团股份有限公司 | Television program recommending method and television program recommending device |
| US20180174218A1 (en) * | 2016-12-19 | 2018-06-21 | Sap Se | Recommendation optmization with a dynamic mixture of frequent and occasional recommendations |
| CN108197285A (en) * | 2018-01-15 | 2018-06-22 | 腾讯科技（深圳）有限公司 | A kind of data recommendation method and device |
| CN108230007A (en) * | 2017-11-28 | 2018-06-29 | 北京三快在线科技有限公司 | A kind of recognition methods of user view, device, electronic equipment and storage medium |
-
2018
- 2018-07-27 CN CN201810843390.7A patent/CN109299994B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103106285A (en) * | 2013-03-04 | 2013-05-15 | 中国信息安全测评中心 | Recommendation algorithm based on information security professional social network platform |
| CN103324750A (en) * | 2013-07-04 | 2013-09-25 | 莫志鹏 | Method for personal screening of photo galleries on the basis of Bayesian network |
| CN105812937A (en) * | 2014-12-30 | 2016-07-27 | Tcl集团股份有限公司 | Television program recommending method and television program recommending device |
| US20180174218A1 (en) * | 2016-12-19 | 2018-06-21 | Sap Se | Recommendation optmization with a dynamic mixture of frequent and occasional recommendations |
| CN108230007A (en) * | 2017-11-28 | 2018-06-29 | 北京三快在线科技有限公司 | A kind of recognition methods of user view, device, electronic equipment and storage medium |
| CN108197285A (en) * | 2018-01-15 | 2018-06-22 | 腾讯科技（深圳）有限公司 | A kind of data recommendation method and device |
Non-Patent Citations (1)
| Title |
|---|
| 崔苹 等: ""基于LDA主题扩展的个性化电影推荐系统"", 《计算机科学与应用》 * |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111800569A (en) * | 2019-04-09 | 2020-10-20 | Oppo广东移动通信有限公司 | Photographing processing method and device, storage medium and electronic equipment |
| CN111800569B (en) * | 2019-04-09 | 2022-02-22 | Oppo广东移动通信有限公司 | Photographing processing method and device, storage medium and electronic equipment |
| CN110866180B (en) * | 2019-10-12 | 2022-07-29 | 平安国际智慧城市科技股份有限公司 | Resource recommendation method, server and storage medium |
| CN110866180A (en) * | 2019-10-12 | 2020-03-06 | 平安国际智慧城市科技股份有限公司 | Resource recommendation method, server and storage medium |
| WO2021082500A1 (en) * | 2019-10-31 | 2021-05-06 | 百果园技术(新加坡)有限公司 | Information recommendation method, apparatus, recommendation server and storage medium |
| CN111080339B (en) * | 2019-11-18 | 2024-01-30 | 口口相传(北京)网络技术有限公司 | Scene-based category preference data generation method and device |
| CN111080339A (en) * | 2019-11-18 | 2020-04-28 | 口口相传(北京)网络技术有限公司 | Method and device for generating category preference data based on scene |
| CN111159393A (en) * | 2019-12-30 | 2020-05-15 | 电子科技大学 | Text generation method for abstracting abstract based on LDA and D2V |
| CN111159393B (en) * | 2019-12-30 | 2023-10-10 | 电子科技大学 | Text generation method for abstract extraction based on LDA and D2V |
| CN111368202A (en) * | 2020-03-06 | 2020-07-03 | 咪咕文化科技有限公司 | Search recommendation method and device, electronic equipment and storage medium |
| CN111368202B (en) * | 2020-03-06 | 2023-09-19 | 咪咕文化科技有限公司 | Search recommendation method and device, electronic equipment and storage medium |
| CN111553748A (en) * | 2020-05-09 | 2020-08-18 | 福州大学 | Android micro-service recommendation method and system based on user scene |
| CN111553748B (en) * | 2020-05-09 | 2022-07-01 | 福州大学 | Android micro-service recommendation method and system based on user scene |
| CN112749343A (en) * | 2021-01-22 | 2021-05-04 | 武汉蔚来能源有限公司 | Resource recommendation method and device and computer storage medium |
| CN113496432A (en) * | 2021-07-06 | 2021-10-12 | 北京爱笔科技有限公司 | Mining method, device and equipment of entity to be recommended and storage medium |
| WO2023019517A1 (en) * | 2021-08-19 | 2023-02-23 | 阿波罗智联(北京)科技有限公司 | Instruction recommendation method and apparatus |
| CN114398486A (en) * | 2022-01-06 | 2022-04-26 | 北京博瑞彤芸科技股份有限公司 | Method and device for intelligently customizing customer acquisition publicity |
| CN114528484A (en) * | 2022-01-26 | 2022-05-24 | 北京金堤科技有限公司 | Preference mining method and device, storage medium and electronic equipment |
| CN114443967A (en) * | 2022-04-08 | 2022-05-06 | 北京并行科技股份有限公司 | Similar application recommendation method, computing device and storage medium |
| CN115470414A (en) * | 2022-11-03 | 2022-12-13 | 安徽商信政通信息技术股份有限公司 | United celebrity recommendation method and recommendation system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109299994B (en) | 2021-09-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109299994A (en) | Recommended method, device, equipment and readable storage medium storing program for executing | |
| CN110008973B (en) | Model training method, method and device for determining target user based on model | |
| CN105302810A (en) | Information search method and apparatus | |
| CN106445963A (en) | Advertisement index keyword automatic generation method and apparatus for APP platform | |
| CN110222285A (en) | The methods of exhibiting of reading page calculates equipment and computer storage medium | |
| CN106445954A (en) | Business object display method and apparatus | |
| CN113592605B (en) | Product recommendation method, device, equipment and storage medium based on similar products | |
| CN106484698A (en) | A kind of method for pushing of search keyword and device | |
| CN110413888A (en) | A kind of books recommended method and device | |
| CN109241455B (en) | Recommended object display method and device | |
| CN109189931A (en) | A kind of screening technique and device of object statement | |
| CN102402717A (en) | Data analysis facility and method | |
| CN113781149A (en) | Information recommendation method and device, computer-readable storage medium and electronic equipment | |
| CN113792134B (en) | User service method and system based on digital twin technology | |
| CN110362662A (en) | Data processing method, device and computer readable storage medium | |
| CN111680213B (en) | Information recommendation method, data processing method and device | |
| CN104077288B (en) | Web page contents recommend method and web page contents recommendation apparatus | |
| CN111767459A (en) | Item recommendation method and device | |
| Aversano et al. | Mining candidate web services from legacy code | |
| CN109558528A (en) | Article method for pushing, device, computer readable storage medium and server | |
| CN112966482A (en) | Report generation method, device and equipment | |
| CN107944589A (en) | The Forecasting Methodology and prediction meanss of ad click rate | |
| CN112749325A (en) | Training method and device for search ranking model, electronic equipment and computer medium | |
| CN114429384B (en) | Intelligent product recommendation method and system based on e-commerce platform | |
| CN110134773A (en) | A kind of search recommended method and system |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |