CN114971817A - Product self-adaptive service method, medium and device based on user demand portrait - Google Patents

Product self-adaptive service method, medium and device based on user demand portrait Download PDF

Info

Publication number
CN114971817A
CN114971817A CN202210906537.9A CN202210906537A CN114971817A CN 114971817 A CN114971817 A CN 114971817A CN 202210906537 A CN202210906537 A CN 202210906537A CN 114971817 A CN114971817 A CN 114971817A
Authority
CN
China
Prior art keywords
product
user
attention
demand
list
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
Application number
CN202210906537.9A
Other languages
Chinese (zh)
Other versions
CN114971817B (en
Inventor
戴礼灿
吴超蓉
宋丹
杨露
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
CETC 10 Research Institute
Original Assignee
CETC 10 Research Institute
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by CETC 10 Research Institute filed Critical CETC 10 Research Institute
Priority to CN202210906537.9A priority Critical patent/CN114971817B/en
Publication of CN114971817A publication Critical patent/CN114971817A/en
Application granted granted Critical
Publication of CN114971817B publication Critical patent/CN114971817B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

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
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Finance (AREA)
  • General Engineering & Computer Science (AREA)
  • Accounting & Taxation (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Computation (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Artificial Intelligence (AREA)
  • Development Economics (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • Strategic Management (AREA)
  • General Business, Economics & Management (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The invention provides a product self-adaptive service method, medium and device based on user demand portrait; the product self-adaptive service method based on the user demand portrait comprises the following steps: generating a user explicit demand portrait through user explicit demand modeling, generating a user implicit demand portrait through user implicit demand modeling, and effectively weighting the user implicit demand portrait and the user implicit demand portrait to form a user personalized demand portrait; starting from two dimensions of accurate recommendation and recall recommendation, flexible parameter configuration schemes under two application scenes are designed, specific parameters are adaptively adjusted according to application requirements of different accuracy rates and recall rates, and the product scene type adaptive service based on the user personalized demand portrait is realized.

Description

Product self-adaptive service method, medium and device based on user demand portrait
Technical Field
The invention relates to the technical field of intelligent recommendation, in particular to a product self-adaptive service method, medium and device based on user demand portrayal.
Background
With the prosperous rise of the network era, the scale of the internet is continuously enlarged, the data resources are explosively increased by the widely distributed internet, and the typical problem of 'information overload' is formed because users are more difficult to obtain information meeting the requirements of the users in a mass information space. The advent of recommendation systems solved this problem well.
Personalized recommendation collects user preferences from a user history record by using technologies such as data mining and the like, and helps the user to acquire interested information so as to solve the problem of overload of internet information. Compared with the traditional information retrieval technology such as a search engine and the like, the recommendation system has the greatest advantages that personalized services can be provided for users, the system guides the users according to personal information or browsing behaviors of the users, and feedback results of the process are different for different users.
But the recommendation system also has the following typical problems:
(1) the problem of "cold start", which is an extreme case of the data sparseness problem, generally occurs when the initial state system score data is too small, and the system has too little available data to make proper recommendations. When a new user appears in the system, because the user information mastered by the system is few, the attention preference of the user cannot be accurately judged, and the targeted information cannot be provided for the user;
(2) checking the service and accurate service problems, defining the product service accuracy as the ratio of the number of products which are interesting (actually needed) by a user in the products pushed by the system to the sum of the number of pushed information products, and reflecting the diversity of the pushed products; the product service recall ratio is defined as the ratio of the number of products which are interesting (actually needed) by a user in the products pushed by the system to the sum of the number of products which are really related to the user requirements in the product library, and the accuracy of pushing the products is reflected. The two indexes have strong correlation, but the recall ratio and the accuracy ratio are basically difficult to be considered simultaneously.
Disclosure of Invention
The present invention is directed to a product adaptive service method, medium and apparatus based on user demand portrayal to solve the above problems.
The invention provides a product self-adaptive service method based on user demand portrait, which comprises the following steps:
step 1: modeling the user explicit requirement to generate a user explicit requirement portrait; specifically, the method comprises the following steps: acquiring product requirements of a user by using a structured requirement modeling tool, wherein the product requirements comprise product topics and product keywords concerned by the user, and outputting a structured user explicit requirement list; the user explicit requirement list is a user explicit requirement portrait;
step 2: modeling the implicit requirements of the user to generate an implicit requirement portrait of the user; specifically, the method comprises the following steps: after a user logs in a product service system, the product service system automatically acquires the interactive operation behaviors and interactive operation product objects of the user, extracts the focus of the user attention by using a demand mining algorithm, and outputs a structured user implicit demand list; the user implicit requirement list is a user implicit requirement portrait;
and step 3: different weights are given to the user explicit demand portrait and the user implicit demand portrait, then the user explicit demand portrait and the user implicit demand portrait are effectively weighted, and a user personalized demand portrait is generated;
and 4, step 4: accurate recommendation and full recommendation of products under two different scenes based on user personalized demand portraits are realized, and specifically:
when the accurate recommendation requirement of the user is higher than the recall recommendation requirement, the product service system automatically recommends the product with high attention to the user on the basis of meeting the preset recall ratio;
when the requirement of the user on the full-checking recommendation is higher than the requirement on the accurate recommendation, the product service system automatically recommends the product with the attention degree not equal to zero to the user on the basis of meeting the preset accuracy rate.
The invention also provides a computer terminal storage medium which stores computer terminal executable instructions used for executing the product self-adaptive service method based on the user demand portrait.
The present invention also provides a computing device comprising:
at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method for product adaptive service based on user demand profiling described above.
In summary, due to the adoption of the technical scheme, the invention has the beneficial effects that:
in order to construct a user personalized demand portrait and update the user personalized demand portrait at regular time, the user explicit demand portrait is generated through user explicit demand modeling, a user implicit demand portrait is generated through user implicit demand modeling, and the user implicit demand portrait are effectively weighted to form the user personalized demand portrait; starting from two dimensions of accurate recommendation and recall recommendation, flexible parameter configuration schemes under two application scenes are designed, specific parameters are adaptively adjusted according to application requirements of different accuracy rates and recall rates, and the product scene type adaptive service based on the user personalized demand portrait is realized.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present invention, and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained according to the drawings without inventive efforts.
FIG. 1 is a flowchart illustrating a method for product adaptive service based on a user demand profile according to an embodiment of the present invention.
FIG. 2 is a flow chart of generating a user explicit request representation according to an embodiment of the present invention.
FIG. 3 is a flowchart illustrating an embodiment of generating a user implicit required representation.
FIG. 4 is a flowchart illustrating the generation of a representation of a user's personalized requirements according to an embodiment of the present invention.
FIG. 5 is a flowchart of product intelligent recommendation in two different scenarios of implementing accurate recommendation and recall recommendation based on a user personalized demand profile in an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. The components of embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations.
Thus, the following detailed description of the embodiments of the present invention, presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Examples
As shown in fig. 1, the present embodiment provides a product adaptive service method based on a user requirement portrait, which includes the following steps:
step 1: modeling the user explicit requirement to generate a user explicit requirement portrait; as shown in fig. 2, that is, using the structured demand modeling tool, collecting product demands of users, including product topics focused by users and product keywords (such as date range, author, source, etc.), and outputting a structured user explicit demand list; the user explicit requirement list is a user explicit requirement representation.
In this embodiment, the user explicit requirement representation is represented as:
Figure 100002_DEST_PATH_IMAGE001
wherein:
Figure 26720DEST_PATH_IMAGE002
representing a user explicit requirements representation;
Figure 100002_DEST_PATH_IMAGE003
an attention list representing a product subject concerned by the user;
Figure 9588DEST_PATH_IMAGE004
representing a user attention list of the product keywords;
T ihist_ (i∈(1~h) Is showing the user's attentioniThe product theme of each product is defined by the number of products,w ihist_ (i∈(1~h) Is showing attention to the useriThe value of interest in the product theme of an individual product,hthe number of product themes for the product that the user is interested in,w hist_1 +w hist_2 +w hist_3 +…+w hhist_ =1；
Kword jhist_ (j∈(1~g) A) represents a product of interest to the userjAn individual product keyword;w’ jhist_ (j∈(1~g) A) represents a product of interest to the userjThe attention of the keywords of each product,gthe number of product keywords for the product of interest to the user,w’ hist_1 +w’ hist_2 +w’ hist_3 +…+w’ ghist_ =1。
step 2: modeling the implicit requirements of the user to generate an implicit requirement portrait of the user; after a user logs in the product service system, the product service system automatically acquires the interactive operation behaviors and interactive operation product objects of the user, extracts the focus of the user attention by using a demand mining algorithm, and outputs a structured user implicit demand list; the user implicit requirement list is the user implicit requirement portrait. As shown in fig. 3, the method specifically includes the following sub-steps:
step 21, collecting and recording the interactive operation behavior and the operation object of the user:
Figure 100002_DEST_PATH_IMAGE005
wherein:
Figure 966043DEST_PATH_IMAGE006
representing a matrix of interactive operational behaviors of a user on a product during login to the product service system,lrepresenting a total number of products having interactive operational behavior including browsing, favorites or downloads,Q i1 represents the user toiThe browsing behavior of the individual products is such that,Q i2 represents the user toiThe act of collecting the individual products is,Q i3 represents the user toiDownload behavior of individual products; for example, the following steps are carried out: when the user is rightiThe product is operatedQ i1 (browsing behavior) when settingQ i1 = 1; if not, then,Q i1 =0。
step 22, calculating the attention of the user to the product:
step 221, constructing a mapping relation matrix of the user operation behavior and the product attention:
Figure 100002_DEST_PATH_IMAGE007
wherein:
Figure 299941DEST_PATH_IMAGE008
a mapping relation matrix representing the user operation behavior and the product attention,w 1 a quantitative value representing the attention of the user's browsing behavior to the product,w 2 a quantitative value representing the product attention of the user's favorite behaviors,w 3 a quantitative value representing the attention of the user to the product;
step 222, calculating a product attention list of the user during logging in the product service system through an interactive operation behavior matrix of the user on the product during logging in the product service system and a mapping relation matrix of the user operation behavior and the product attention:
Figure 100002_DEST_PATH_IMAGE009
wherein:
Figure 124678DEST_PATH_IMAGE010
indicating that the user is interested in the product during the login to the product service system,D i (i∈(1~l) Represents the user is rightiA value of interest for the individual product;
step 23, calculating a user implicit requirement list, and generating a user implicit requirement portrait:
step 231, collecting product themes and product keywords for each product concerned by the user, and outputting a product theme list and a product keyword list:
the product theme list is represented as:
Figure 712785DEST_PATH_IMAGE011
wherein:T i (i∈(1~l) Is expressed asiProduct subject of individual product, one product is considered to have and only one product subject;
the product keyword list is represented as:
Figure 100002_DEST_PATH_IMAGE012
wherein: [Kword i ](i∈(1~l) Is expressed asiProduct keywords for each product, wherein a product is considered to have one or more product keywords; first, theiThe product keyword list for an individual product is represented as:
Kword i =[[Kword i1 ,w i1 ],[Kword i2 ,w i2 ],…,[Kword ik ,w ik ]]，i∈(1~l)；
wherein:kis shown asiThe number of product keywords for each product; [Kword iq ,w iq ](q∈(1~k) In (c) in (c),Kword iq is shown asiThe first productqThe key words of each product are set,w iq is shown asqWeight of each product keyword;
step 232, calculating the attention of the user to the product theme, and if one product is considered to have one and only one product theme, the user is interested in the second product themeiProduct theme attention of individual product, namely user toiAttention of individual products:
firstly, calculating the product theme attention of a user to each product, and expressing as follows:
Figure 433617DEST_PATH_IMAGE013
wherein the content of the first and second substances,
Figure 989232DEST_PATH_IMAGE014
(i∈(1~l) Represents the user is rightiProduct theme of individual productT i The attention is
Figure 100002_DEST_PATH_IMAGE015
，D i For the user toiThe attention of the individual product;
then, a union set is obtained for the product theme attention degree list, the same theme is merged, and the attention degree of the user to the product theme is output: when the product themes of a plurality of products are the same, the attention degree of the user to the product theme is the sum of all attention degrees under the product theme; i.e. when it comes toiA product and the secondjThe product theme of each product is the same, i.e.T i =T j (i∈(1~l),j∈(1~l),i≠j) The user is on the theme of the productT i Degree of attention of
Figure 227446DEST_PATH_IMAGE016
(ii) a By combining a plurality of attention degrees under the same product theme, outputting a product theme attention degree list of a user, and showing:
Figure 100002_DEST_PATH_IMAGE017
wherein:
Figure 325852DEST_PATH_IMAGE018
a list representing the user's attention to the subject of the product;
Figure 100002_DEST_PATH_IMAGE019
(i∈(1~n)，n≤l) Represents the user toiProduct theme of individual productT i Has a value of interest of
Figure 434946DEST_PATH_IMAGE015
；
And finally, carrying out normalization processing on the attention of each product theme, wherein a normalization formula is as follows:
Figure 228590DEST_PATH_IMAGE020
outputting a normalized list of the user's attention to the product theme, expressed as:
Figure 100002_DEST_PATH_IMAGE021
wherein the content of the first and second substances,w i (i∈(1~n) Is normalized user pair numberiThe product theme awareness of an individual product,w 1 +w 2 +…+w n =1；
step 233, calculating the attention of the user to the product keywords, considering that one product has one or more product keywords, and the attention of the user to the product keywords is the product attention multiplied by the product keyword weight:
firstly, calculating the attention degree of a user to each product keyword, and expressing the attention degree as follows:
Figure 395129DEST_PATH_IMAGE022
wherein:
Figure 100002_DEST_PATH_IMAGE023
represents the user toiA list of the attention of the product keywords in each product;kis shown asiThe number of product keywords for each product; [Kword ik ,w ik ]Is shown asiThe first productkThe product keywords and the attention degree of the user to the product keywords after considering the product attention degree;
then, a union set is obtained for the product keyword attention degree list, the same product keywords are merged, and the attention degree of the user to the product keywords is output: traversing user attention list to product keywords
Figure 207096DEST_PATH_IMAGE024
When a plurality of product keywords are the same, the attention of the user to the product keywords is the sum of all the attention of the user to the product keywords; i.e. when it comes toiProduct key words of individual productKword ik And a first step ofjProduct key words of individual productKword jm Are identical, i.e. thatKword ik =Kword jm The user can search the product key wordsKword ik Has a degree of attention ofw ik =w ik +w jm (ii) a Outputting a product keyword attention list of a user by combining a plurality of attention degrees under the same product keyword, and showing:
Figure 100002_DEST_PATH_IMAGE025
wherein:
Figure 676255DEST_PATH_IMAGE026
representing a product keyword attention degree list of a user;
Figure 550670DEST_PATH_IMAGE027
(i∈(1~p) Represents the user is rightiProduct key wordKword i Has a value of interest of
Figure 100002_DEST_PATH_IMAGE028
；
And finally, carrying out normalization processing on the attention of each product keyword, wherein a normalization formula is as follows:
Figure 645534DEST_PATH_IMAGE029
outputting a normalized list of the attention of the user to the product keywords, wherein the list is expressed as:
Figure DEST_PATH_IMAGE030
wherein the content of the first and second substances,w i (i∈(1~p) Is normalized toiThe degree of attention of the keywords of each product,w 1 +w 2 +…+ w p =1；
step 234, generating a user implicit requirement list, namely a user implicit requirement portrait, based on the user product theme attention and the user product keyword attention calculated in steps 232 and 233, and showing as:
Figure 328319DEST_PATH_IMAGE031
wherein:
Figure 561854DEST_PATH_IMAGE032
and the list representing the implicit requirements of the user consists of the user product subject attention list and the user product keyword attention list which are calculated in the steps 232 and 233.
And step 3: different weights are given to the user explicit demand portrait and the user implicit demand portrait, then the user explicit demand portrait and the user implicit demand portrait are effectively weighted, and a user personalized demand portrait is generated; as shown in fig. 4, the method specifically includes the following sub-steps:
step 31, distributing weights for the user historical demand portrait and the user implicit demand portrait:
w hist +w inter =1
wherein the content of the first and second substances,w hist a weight representing a representation of the historical demand of the user,w inter weights representing a user implicit demand profile;
during the first time that the user logs in the product service system, the user historical demand portrait is the user explicit demand portrait; when the user logs in the product service system again, the user historical demand portrait is a user personalized demand portrait generated during the previous log-in period of the product service system;
step 32, calculating the product theme attention of the personalized demand:
step 321, weighting the product theme attention degree in the user historical demand portrait and the product theme attention degree in the user implicit demand portrait by using the weight output in step 31, and outputting the product theme attention degree of the user personalized demand, which is expressed as:
Figure 189144DEST_PATH_IMAGE033
step 322, merging the same product topics in the product topic attention list: traversing user attention list of product subject
Figure 100002_DEST_PATH_IMAGE034
When a plurality of product themes are the same, the attention of the user to the product theme is the sum of all the attention to the product theme; that is, when the product theme T hist_1 With product theme T n When the same, the user is on productionTheme T hist_1 Has a degree of attention of (w hist ×w hist_1 +w inter ×w n ) (ii) a By combining a plurality of attention degrees under the same product theme, a product theme attention degree list of the user personalized demand list is output, and is expressed as follows:
Figure 822120DEST_PATH_IMAGE035
wherein:
Figure 100002_DEST_PATH_IMAGE036
respectively representing the product theme concerned by the user and the attention value of the product theme;
step 33, calculating the product keyword attention of the personalized demand:
step 331, weighting the product keyword attention in the user historical demand image and the product keyword attention in the user implicit demand image by using the weighted value output in step 31, and outputting the product keyword attention of the user personalized demand, which is expressed as:
Figure 359412DEST_PATH_IMAGE037
step 332, merging the similar product keywords in the product keyword list:
setting similarity threshold of product keywordsxIf the similarity of the two product keywords is greater thanxIf the two product keywords are similar words, merging can be carried out;
traversing the product keywords of the product keyword attention list of the user, calculating the word vector of each product keyword by using a word vector algorithm, calculating the cosine value of the included angle between the word vectors of every two product keywords by using a cosine similarity algorithm, and if the cosine value is larger than the similarity threshold valuexIf the two product keywords are similar, merging can be carried out, and the attention degree of the two product keywords is the sum of the attention degrees of the product keywords;
for example, the following steps are carried out: setting a similarity thresholdx=0.9, calculating the product key words by traversing the product key words, and calculating the word vector and the cosine similarityKword 1 AndKword g+m is 0.95, it is considered thatKword 1 AndKword g+m similarly, the product key wordKword 1 AndKword g+m are combined intoKword 1 With a focus value of: (w 1 +w g m+ ) (ii) a Outputting a product keyword attention list of the user personalized requirements, wherein the product keyword attention list is expressed as:
Figure 100002_DEST_PATH_IMAGE038
wherein:
Figure 422570DEST_PATH_IMAGE039
。
and 4, step 4: accurate recommendation and full recommendation of products under two different scenes based on user personalized demand portrait are achieved; as shown in fig. 5, the method specifically includes the following sub-steps:
step 41, calculating the attention degree of the user to the product based on the personalized demand product theme:
step 411, traverse all products in the product library, perform product theme acquisition on each product in the product library, and output the product theme of the ith productT i ；
Step 412, traverse the user personalized demand topic list
Figure 100002_DEST_PATH_IMAGE040
Subject the productT i With the product theme in the user's personalized demand theme listT j And (3) performing matching calculation: if the matching is successful, stopping traversing and comparing the first step with the second stepiAttention of individual productD T_i Equal to the product theme in the user personalized demand theme listT j Degree of attention ofw j (ii) a If it isIf the configuration fails, go to step 413;
in step 413, the process repeats step 412,j=j+1,j≤nup to a list of themes of the demand personalized to the user
Figure 802736DEST_PATH_IMAGE040
Finishing traversing; to show thatiProduct theme of individual productT i If the matching with all the product themes in the user personalized demand theme list fails, the attention degree of the user to the ith product based on the personalized demand themew i =0；
Step 414, repeating steps 411-413 until all products in the product library are traversed, and outputting the attention degree of the user to the products based on the user personalized demand theme, wherein the attention degree is expressed as:
Figure 193397DEST_PATH_IMAGE041
wherein:yrepresenting the total number of products in the product library;D T_i (i∈(1~y) Representing user pairs based on user personalized requirements topicsiThe attention of the individual product;
and 42, calculating the attention degree of the user to the product based on the personalized demand product keywords:
step 421, go through all the products in the product library, collect the product key words for each product in the product library, and output the firstiA product keyword list for each product, represented as:
{Kword i }=[[Kword i1 ,w i1 ],[Kword i2 ,w i2 ],…,[Kword ik ,w ik ]]
wherein: {Kword i Means the firstiA product keyword list of individual products; [Kword ij ,w ij ](j∈(1~k) Respectively representiThe first productjProduct key word and method for obtaining the key wordThe weight of the weight is calculated,kis shown asiThe number of product keywords for each product;
step 422, respectively calculate the secondiFeature vectors of the keyword list of the individual product and the user personalized demand product:
firstly to the firstiThe product keyword list of each product and the product keyword list of the user personalized demand are merged, and a bag-of-words model is output and expressed as:
Figure 100002_DEST_PATH_IMAGE042
wherein the content of the first and second substances,Kword i1 ,Kword i2 ,…,Kword ik respectively representiOf a productkThe key words of each product are set,Kword 1 ,Kword 2 ,…,Kword f f product keywords in the product keyword list of the user personalized demand are respectively;
for example, the following steps are carried out: when the product keywordKword i1 =Kword 1 Then, the output bag-of-words model is:
Figure 975408DEST_PATH_IMAGE043
bag of words modelUThe product key words in (1) are indexes, the corresponding numerical values are represented by product key word weights, if the product key words do not appear in the product key word list, the weights are 0, and the first step is respectively generatediThe word bag vectors of the keywords of the individual products and the personalized demand products of the users are called feature vectors.
For example, the following steps are carried out: when the product keywordKword i1 =Kword 1 Then output the firstiThe bag-of-words vectors for individual products and the bag-of-words vectors for the keyword list of the user-customized demand product are as follows:
Figure 100002_DEST_PATH_IMAGE044
wherein:Kword i1 ,Kword i2 ,…,Kword ik ,Kword 2 ,…,Kword f is an index number; [w i1 ,w i2 ,…,w ik ,0,0]Is as followsiFeature vectors of individual products; [w i ,0,…,0,w 2 ,w f ]And personalizing the feature vector of the product keyword list for the user.
Step 423, calculating the product attention of the user based on the personalized demand product keyword:
using a cosine calculation formulaiCosine value of included angle between feature vector of individual product and feature vector of keyword list of user personalized demand productiSimilarity between the product and the keyword list of the user personalized demand product; the closer the cosine value of the included angle of the feature vectors is to 1, the higher the similarity is, the higher the attention of the user to the product is, and the calculation formula is as follows:
Figure 206538DEST_PATH_IMAGE045
wherein:D Kword_i for the user pair based on the personalized demand product keywordiThe attention of the individual product;
step 424, repeating steps 421-423 until all products in the product library are traversed, and outputting the user attention to the products based on the user personalized demand product keywords, which is expressed as:
Figure 100002_DEST_PATH_IMAGE046
wherein:yrepresenting the total number of products in the product library;D Kword_i (i∈(1~y) User pair representing keywords for a product based on user personalized requirementsiThe attention of the individual product;
and 43, calculating the product attention based on the personalized demand portrait based on the output results of the step 41 and the step 42:
step 431, traversing all products in the product library, and based on the user pair of the personalized demand portraitiThe attention degree of each product is based on the user pair of the personalized demand product keywordsiAttention of individual product and user pair based on personalized demand product themeiThe product of the attention weights of the individual products is expressed as:
Figure 949366DEST_PATH_IMAGE047
wherein:
Figure DEST_PATH_IMAGE048
representing user pairs based on personalized demand figuresiThe attention of the individual product;D Kword_i representing user pairs based on personalized demand product keywordsiThe attention of the individual product;D T_i representing pairs of users based on personalized requirements topicsiThe attention of the individual product; max (D T_1 ,D T_2 ,…,D T_y ) Representing the maximum value of the attention degree of the user to the product based on the personalized demand product theme; min (D T_1 ,D T_2 ,…,D T_y ) Representing the minimum value of the attention degree of the user to the product based on the personalized demand product theme;
step 432, after the traversal is finished, outputting a list of the attention degrees of the user to all the products in the product library, wherein the list is represented as:
Figure 924145DEST_PATH_IMAGE049
wherein:
Figure 295083DEST_PATH_IMAGE050
the attention degree list of the user to all products in the product library is represented;yrepresenting the number of products in the product library;D i (i∈(1~y) Representing user pairs based on personalized demand figuresiThe attention of the individual product;
and step 44, realizing accurate recommendation and full recommendation based on the user personalized demand portrait based on the product attention under two different scenes:
when the accurate recommendation requirement of the user is higher than the recall recommendation requirement of the user, the product service system automatically recommends the product with high attention to the user on the basis of meeting the preset recall ratio;
when the requirement of the user on the full-checking recommendation is higher than the requirement on the accurate recommendation, the product service system automatically recommends the product with the attention degree not equal to zero to the user on the basis of meeting the preset accuracy rate.
Therefore, the scene type product self-adaptive service based on the user personalized demand portrait is realized.
Further, in some embodiments, a computer terminal storage medium is provided storing computer terminal executable instructions for performing a product adaptive service method based on user demand profiling as described in the previous embodiments. Examples of the computer storage medium include a magnetic storage medium (e.g., a floppy disk, a hard disk, etc.), an optical recording medium (e.g., a CD-ROM, a DVD, etc.), or a memory such as a memory card, a ROM, a RAM, or the like. The computer storage media may also be distributed over a network-connected computer system, such as an application store.
Furthermore, in some embodiments, a computing device is presented, comprising: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform a method for product adaptive services based on user demand profiling as described in the previous embodiments. Examples of computing devices include PCs, tablets, smart phones, or PDAs, among others.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (10)

1. A product self-adaptive service method based on user demand portrait is characterized by comprising the following steps:
step 1: modeling the user explicit requirement to generate a user explicit requirement portrait;
step 2: modeling the implicit requirements of the user to generate a user implicit requirement portrait;
and step 3: different weights are given to the user explicit demand portrait and the user implicit demand portrait, then the user explicit demand portrait and the user implicit demand portrait are effectively weighted, and a user personalized demand portrait is generated;
and 4, step 4: accurate recommendation and comprehensive recommendation of products under two different scenes based on user personalized demand portraits are achieved.
2. The method for product adaptive services based on user demand representation of claim 1, wherein the step 1 of generating user explicit demand representation by user explicit demand modeling comprises:
acquiring product requirements of a user by using a structured requirement modeling tool, wherein the product requirements comprise product topics and product keywords concerned by the user, and outputting a structured user explicit requirement list; the user explicit requirement list is a user explicit requirement representation.
3. The user demand representation-based product adaptive service method of claim 2, wherein the user explicit demand representation is represented as:
Figure DEST_PATH_IMAGE001
wherein:
Figure 583839DEST_PATH_IMAGE002
representing a user explicit requirements representation;
Figure DEST_PATH_IMAGE003
an attention list representing a product subject concerned by the user;
Figure 327673DEST_PATH_IMAGE004
representing a user attention list of the product keywords;T ihist_ (i∈(1~h) Is showing the user's attentioniThe product theme of each product is defined by the number of products,w ihist_ (i∈(1~h) Is showing attention to the useriThe degree of concern about the product theme of an individual product,hthe number of product themes for the product that the user is interested in,w hist_1 +w hist_2 +w hist_3 +…+w hhist_ =1；Kword jhist_ (j∈(1~g) Second to represent products of interest to the userjAn individual product keyword;w’ jhist_ (j∈(1~g) A) represents a product of interest to the userjThe attention of the keywords of each product,gthe number of product keywords for the product of interest to the user,w’ hist_1 +w’ hist_2 +w’ hist_3 +…+w’ ghist_ =1。
4. the user demand portrayal-based product adaptive service method according to claim 1, wherein the user implicit demand modeling generation method of the user implicit demand portrayal in step 2 comprises:
after a user logs in a product service system, the product service system automatically acquires the interactive operation behaviors and interactive operation product objects of the user, extracts the focus of the user attention by using a demand mining algorithm, and outputs a structured user implicit demand list; the user implicit requirement list is the user implicit requirement portrait.
5. The user demand portrayal-based product adaptation service method of claim 4, wherein the step 2 of generating the user implicit demand portrayal by user implicit demand modeling comprises the following sub-steps:
step 21, collecting and recording the interactive operation behavior and the operation object of the user:
Figure DEST_PATH_IMAGE005
wherein:
Figure 266810DEST_PATH_IMAGE006
representing a matrix of interactive operational behaviors of a user on a product during login to the product service system,lrepresenting the total number of products having interactive operational behavior including browsing, favorites or downloads,Q i1 represents the user toiThe browsing behavior of the individual products is such that,Q i2 represents the user toiThe act of collecting the individual products is,Q i3 represents the user toiDownload behavior of individual products;
step 22, calculating the attention of the user to the product:
step 221, constructing a mapping relation matrix of the user operation behavior and the product attention:
Figure DEST_PATH_IMAGE007
wherein:
Figure 262448DEST_PATH_IMAGE008
a mapping relation matrix representing the user operation behavior and the product attention,w 1 a quantitative value representing the attention of the user's browsing behavior to the product,w 2 a quantitative value representing the product attention of the user's favorite behaviors,w 3 representing download behavior of a userA quantified value for product attention;
step 222, calculating a product attention list of the user during logging in the product service system through an interactive operation behavior matrix of the user on the product during logging in the product service system and a mapping relation matrix of the user operation behavior and the product attention:
Figure DEST_PATH_IMAGE009
wherein:
Figure 109488DEST_PATH_IMAGE010
indicating that the user is interested in the product during the login to the product service system,D i (i=1~l) Represents the user toiA value of interest for the individual product;
step 23, calculating a user implicit requirement list, and generating a user implicit requirement portrait:
step 231, collecting product themes and product keywords for each product concerned by the user, and outputting a product theme list and a product keyword list:
the product theme list is represented as:
Figure 243797DEST_PATH_IMAGE011
wherein:T i (i∈(1~l) Denotes to the firstiProduct subject of individual product, one product is considered to have and only one product subject;
the product keyword list is represented as:
Figure DEST_PATH_IMAGE012
wherein: [Kword i ](i∈(1~l) Is expressed asiProduct key word of each product, and considering a productThere are one or more product keywords; first, theiThe product keyword list for an individual product is represented as:
Kword i =[[Kword i1 ,w i1 ],[Kword i2 ,w i2 ],…,[Kword ik ,w ik ]]，i=1~l；
wherein:kdenotes the firstiThe number of product keywords for each product; [Kword iq ,w iq ](q=1~k) In (1),Kword iq denotes the firstiThe first productqThe key words of each product are set,w iq is shown asqWeight of each product keyword;
step 232, calculating the attention of the user to the product theme, and if one product is considered to have one and only one product theme, the user is interested in the second product themeiProduct theme attention of individual product, namely user toiAttention of individual products:
firstly, calculating the product theme attention of a user to each product, and expressing as follows:
Figure 529285DEST_PATH_IMAGE013
wherein the content of the first and second substances,
Figure 922089DEST_PATH_IMAGE014
(i∈(1~l) Represents the user is rightiProduct theme of individual productT i The attention is
Figure DEST_PATH_IMAGE015
，D i For the user toiThe attention of the individual product;
then, a union set is obtained for the product theme attention degree list, the same theme is merged, and the attention degree of the user to the product theme is output: when the product themes of a plurality of products are the same, the attention of the user to the product theme is the sum of all the attention under the product theme; by combining a plurality of attention degrees under the same product theme, outputting a product theme attention degree list of a user, and showing:
Figure 383158DEST_PATH_IMAGE016
wherein:
Figure DEST_PATH_IMAGE017
a list representing the user's attention to the subject of the product;
Figure 609740DEST_PATH_IMAGE018
(i∈(1~n)，n≤l) Represents the user toiProduct theme of individual productT i Has a value of interest of
Figure 241578DEST_PATH_IMAGE015
；
And finally, carrying out normalization processing on the attention of each product theme, wherein a normalization formula is as follows:
Figure DEST_PATH_IMAGE019
outputting a normalized list of the user's attention to the product theme, expressed as:
Figure 313439DEST_PATH_IMAGE020
wherein the content of the first and second substances,w i (i∈(1~n) Is normalized user pair numberiThe product theme awareness of an individual product,w 1 +w 2 +…+ w n =1；
step 233, calculating the attention of the user to the product keywords, considering that one product has one or more product keywords, and the attention of the user to the product keywords is the product attention multiplied by the product keyword weight:
firstly, calculating the attention degree of a user to each product keyword, and expressing the attention degree as follows:
Figure DEST_PATH_IMAGE021
wherein:
Figure 629014DEST_PATH_IMAGE022
represents the user toiA list of the attention of the product keywords in each product;kis shown asiThe number of product keywords for each product; [Kword ik ,w ik ]Is shown asiThe first product ofkThe product keywords and the attention degree of the user to the product keywords after considering the product attention degree;
then, a union set is obtained for the product keyword attention degree list, the same product keywords are merged, and the attention degree of the user to the product keywords is output: traversing user attention list to product keywords
Figure DEST_PATH_IMAGE023
When a plurality of product keywords are the same, the attention of the user to the product keywords is the sum of all the attention of the user to the product keywords; i.e. when it comes toiProduct key words of individual productKword ik And a firstjProduct key words of individual productKword jm Are identical, i.e. thatKword ik =Kword jm The user can search the product key wordsKword ik Has a degree of attention ofw ik =w ik +w jm (ii) a Outputting a product keyword attention list of a user by combining a plurality of attention degrees under the same product keyword, and showing:
Figure 151131DEST_PATH_IMAGE024
wherein:
Figure DEST_PATH_IMAGE025
representing a product keyword attention degree list of a user;
Figure 552157DEST_PATH_IMAGE026
(i∈(1~p) Represents the user is rightiProduct key wordKword i Has a value of interest of
Figure 896550DEST_PATH_IMAGE027
；
And finally, carrying out normalization processing on the attention of each product keyword, wherein the normalization formula is as follows:
Figure DEST_PATH_IMAGE028
outputting a normalized list of the attention of the user to the product keywords, wherein the list is expressed as:
Figure 318829DEST_PATH_IMAGE029
wherein the content of the first and second substances,w i (i∈(1~p) Is normalized toiThe degree of attention of the keywords of each product,w 1 +w 2 +…+w p =1；
step 234, generating a user implicit requirement list, namely a user implicit requirement portrait, based on the user product theme attention and the user product keyword attention calculated in steps 232 and 233, and showing as:
Figure 621634DEST_PATH_IMAGE030
wherein:
Figure 713218DEST_PATH_IMAGE031
and the list representing the implicit requirements of the user consists of the user product subject attention list and the user product keyword attention list which are calculated in the steps 232 and 233.
6. The method for product adaptive services based on user demand portraits as claimed in claim 5, wherein the step 3 of assigning different weights to the user explicit demand portraits and the user implicit demand portraits, and then effectively weighting the user explicit demand portraits and the user implicit demand portraits to generate the user personalized demand portraits comprises the sub-steps of:
step 31, distributing weights for the user historical demand portrait and the user implicit demand portrait:
w hist +w inter =1
wherein the content of the first and second substances,w hist a weight representing a representation of the historical demand of the user,w inter weights representing a user implicit demand profile;
step 32, calculating the product theme attention of the personalized demand:
step 321, weighting the product theme attention degree in the user historical demand portrait and the product theme attention degree in the user implicit demand portrait by using the weight output in step 31, and outputting the product theme attention degree of the user personalized demand, which is expressed as:
Figure DEST_PATH_IMAGE032
step 322, merging the same product topics in the product topic attention list: traversing user attention list of product subject
Figure 392461DEST_PATH_IMAGE033
When a plurality of product themes are the same, the attention of the user to the product theme is the sum of all the attention to the product theme; that is, when the product theme T hist_1 With product theme T n When the same, the user is on the product theme T hist_1 Has a degree of attention of (w hist ×w hist_1 +w inter ×w n ) (ii) a By combining a plurality of attention degrees under the same product theme, a product theme attention degree list of the user personalized demand list is output, and is expressed as follows:
Figure DEST_PATH_IMAGE034
wherein:
Figure 666317DEST_PATH_IMAGE035
respectively representing the product theme concerned by the user and the attention value of the product theme;
step 33, calculating the attention of the product keywords of the personalized requirements:
step 331, weighting the product keyword attention in the user historical demand image and the product keyword attention in the user implicit demand image by using the weighted value output in step 31, and outputting the product keyword attention of the user personalized demand, which is expressed as:
Figure DEST_PATH_IMAGE036
step 332, merging the similar product keywords in the product keyword list:
setting similarity threshold of product keywordsxIf the similarity of the two product keywords is greater thanxIf the two product keywords are similar words, merging can be carried out;
traversing the product keywords of the user to the product keyword attention degree list, and counting by using a word vector algorithmCalculating word vectors of each product keyword, calculating cosine values of included angles between the word vectors of every two product keywords by using a cosine similarity calculation method, and if the cosine values are larger than a similarity threshold valuexAnd if so, considering that the two product keywords are similar and can be merged, wherein the attention degree of the product keywords is the sum of the attention degrees of the product keywords.
7. The method for product adaptive service based on user demand representation of claim 6, wherein in step 31, the user historical demand representation is the user explicit demand representation during the initial login of the user to the product service system; when the user logs in the product service system again, the user historical demand portrait is the user personalized demand portrait generated during the previous log-in period of the product service system.
8. The user demand profile-based product adaptive service method according to claim 1, wherein the method for realizing accurate recommendation and full recommendation based on the user personalized demand profile in step 4 comprises the following sub-steps:
step 41, calculating the attention degree of the user to the product based on the personalized demand product theme:
step 411, traverse all products in the product library, perform product theme acquisition on each product in the product library, and output the product theme of the ith productT i ；
Step 412, traverse the user personalized demand topic list
Figure 280969DEST_PATH_IMAGE037
Subject the productT i With the product theme in the user's personalized demand theme listT j And (3) performing matching calculation: if the matching is successful, stopping traversing and comparing the first step with the second stepiAttention of individual productD T_i Equal to the product theme in the user personalized demand theme listT j Degree of attention ofw j (ii) a If the matching fails, go to step 413;
in step 413, the process repeats step 412,j=j+1,j≤nup to a list of themes of the demand personalized to the user
Figure 250062DEST_PATH_IMAGE037
Ending traversal; to show thatiProduct theme of individual productT i If the user's personal demand theme list fails to match all the product themes, the user's attention to the ith product based on the personal demand themew i =0；
Step 414, repeating steps 411-413 until all products in the product library are traversed, and outputting the attention degree of the user to the products based on the user personalized demand theme, wherein the attention degree is expressed as:
Figure DEST_PATH_IMAGE038
wherein:yrepresenting the total number of products in the product library;D T_i (i∈(1~y) Representing user pairs based on user personalized requirements topicsiThe attention of the individual product;
step 42, calculating the attention degree of the user to the product based on the personalized demand product keywords:
step 421, go through all the products in the product library, collect the product key words for each product in the product library, and output the firstiA product keyword list for each product, represented as:
{Kword i }=[[Kword i1 ,w i1 ],[Kword i2 ,w i2 ],…,[Kword ik ,w ik ]]
wherein: {Kword i Means the firstiA product keyword list of individual products; [Kword ij ,w ij ](j∈(1~k) Respectively representiThe first product ofjProduct key word and the keyThe weight of the key-word or words,kis shown asiThe number of product keywords for each product;
step 422, respectively calculate theiFeature vectors of the keyword list of the individual product and the user personalized demand product:
firstly to the firstiThe product keyword list of each product and the product keyword list of the user personalized demand are merged, and a bag-of-words model is output and expressed as:
Figure 592050DEST_PATH_IMAGE039
wherein the content of the first and second substances,Kword i1 ,Kword i2 ,…,Kword ik respectively representiOf a productkThe key words of each product are set,Kword 1 ,Kword 2 ,…,Kword f f product keywords in the product keyword list of the user personalized demand are respectively;
for example, the following steps are carried out: when the product keywordKword i1 =Kword 1 Then, the output bag-of-words model is:
Figure DEST_PATH_IMAGE040
bag of words modelUThe product key words in (1) are indexes, the corresponding numerical values are represented by product key word weights, if the product key words do not appear in the product key word list, the weights are 0, and the first step is respectively generatediThe word bag vectors of the keywords of the individual products and the user personalized demand products are called as feature vectors;
step 423, calculating the product attention of the user based on the personalized demand product keyword:
using a cosine calculation formulaiCosine value of included angle between feature vector of individual product and feature vector of keyword list of user personalized demand productiSimilarity between the product and the keyword list of the user personalized demand product; meterThe calculation formula is as follows:
Figure 471145DEST_PATH_IMAGE041
wherein:D Kword_i for the user pair based on the personalized demand product keywordiThe attention of the individual product;
step 424, repeating steps 421-423 until all products in the product library are traversed, and outputting the user attention to the products based on the user personalized demand product keywords, which is expressed as:
Figure DEST_PATH_IMAGE042
wherein:yrepresenting the total number of products in the product library;D Kword_i (i∈(1~y) User pair representing keywords for a product based on user personalized requirementsiThe attention of the individual product;
and 43, calculating the product attention based on the personalized demand portrait based on the output results of the step 41 and the step 42:
step 431, traversing all products in the product library, and based on the user pair of the personalized demand portraitiThe attention of each product is based on the user pair of the keyword of the personalized demand productiAttention of individual product and user pair based on personalized demand product themeiThe product of the attention weights of the individual products is expressed as:
Figure 381332DEST_PATH_IMAGE043
wherein:
Figure DEST_PATH_IMAGE044
representing user pairs based on personalized demand figuresiThe attention of the individual product;D Kword_i representing products based on personalized demandUser pairing of article keywordsiThe attention of the individual product;D T_i representing pairs of users based on personalized requirements topicsiThe attention of the individual product; max (D T_1 ,D T_2 ,…,D T_y ) Representing the maximum value of the product attention of the user based on the personalized demand product theme; min (D T_1 ,D T_2 ,…,D T_y ) Representing the minimum value of the attention degree of the user to the product based on the personalized demand product theme;
step 432, after the traversal is finished, outputting a list of the attention degrees of the user to all the products in the product library, wherein the list is represented as:
Figure 227934DEST_PATH_IMAGE045
wherein:
Figure DEST_PATH_IMAGE046
the attention degree list of the user to all products in the product library is represented;yrepresenting the number of products in the product library;D i (i∈(1~y) Representing user pairs based on personalized demand figuresiThe attention of the individual product;
and step 44, realizing accurate recommendation and full recommendation based on the user personalized demand portrait based on the product attention under two different scenes:
when the accurate recommendation requirement of the user is higher than the recall recommendation requirement, the product service system automatically recommends the product with high attention to the user on the basis of meeting the preset recall ratio;
when the requirement of the user on the recall recommendation is higher than that on the accurate recommendation, the product service system automatically recommends the products with the attention degree not equal to zero to the user on the basis of meeting the preset accuracy rate.
9. A computer terminal storage medium storing computer terminal executable instructions for performing the method for serving a product adaptively according to a user demand profile as claimed in any one of claims 1 to 8.
10. A computing device, comprising:
at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to cause the at least one processor to perform the method of product adaptive service based on user demand portrayal as claimed in any one of claims 1-8.
CN202210906537.9A 2022-07-29 2022-07-29 Product self-adaptive service method, medium and device based on user demand portrait Active CN114971817B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210906537.9A CN114971817B (en) 2022-07-29 2022-07-29 Product self-adaptive service method, medium and device based on user demand portrait

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210906537.9A CN114971817B (en) 2022-07-29 2022-07-29 Product self-adaptive service method, medium and device based on user demand portrait

Publications (2)

Publication Number Publication Date
CN114971817A true CN114971817A (en) 2022-08-30
CN114971817B CN114971817B (en) 2022-11-22

Family

ID=82968879

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210906537.9A Active CN114971817B (en) 2022-07-29 2022-07-29 Product self-adaptive service method, medium and device based on user demand portrait

Country Status (1)

Country Link
CN (1) CN114971817B (en)

Citations (20)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110307813A1 (en) * 2010-06-11 2011-12-15 International Business Machines Corporation Interactive Ring-Shaped Interface
US20120066072A1 (en) * 2007-05-25 2012-03-15 Issar Amit Kanigsberg Recommendation Systems and Methods Using Interest Correlation
CN103235824A (en) * 2013-05-06 2013-08-07 上海河广信息科技有限公司 Method and system for determining web page texts users interested in according to browsed web pages
CN103235823A (en) * 2013-05-06 2013-08-07 上海河广信息科技有限公司 Method and system for determining current interest of users according to related web pages and current behaviors
US20140101243A1 (en) * 2012-10-05 2014-04-10 Facebook, Inc. Method and apparatus for identifying common interest between social network users
US20150120722A1 (en) * 2013-10-31 2015-04-30 Telefonica Digital Espana, S.L.U. Method and system for providing multimedia content recommendations
CN105142028A (en) * 2015-07-29 2015-12-09 华中科技大学 Television program content searching and recommending method oriented to integration of three networks
CN105677846A (en) * 2016-01-06 2016-06-15 中国传媒大学 Recommendation system and construction method thereof
WO2017193323A1 (en) * 2016-05-12 2017-11-16 深圳大学 User preference-based personalized recommendation method and system utilizing same
US20180197090A1 (en) * 2017-01-10 2018-07-12 Facebook, Inc. Externally informed counterfactual prediction
US10394885B1 (en) * 2016-03-15 2019-08-27 Intuit Inc. Methods, systems and computer program products for generating personalized financial podcasts
CN110263106A (en) * 2019-06-25 2019-09-20 中国人民解放军国防科技大学 Collaborative public opinion fraud detection method and device
CN110362740A (en) * 2019-06-10 2019-10-22 河海大学 A kind of water conservancy gateway information mixed recommendation method
CN111079028A (en) * 2019-12-04 2020-04-28 上海财经大学 Collaborative filtering recommendation system and method based on multi-source auxiliary information
CN112269932A (en) * 2020-10-30 2021-01-26 金天国际医疗科技有限公司 Big data-based small and medium enterprise resource integration processing system
CN112632397A (en) * 2021-01-04 2021-04-09 同方知网(北京)技术有限公司 Personalized recommendation method based on multi-type academic achievement portrait and mixed recommendation strategy
WO2021070681A1 (en) * 2019-10-10 2021-04-15 株式会社村田製作所 Degree-of-interest evaluation system and degree-of-interest evaluation method
CN113988060A (en) * 2021-08-31 2022-01-28 中国电子科技集团公司第十五研究所 Data information recommendation method and device, terminal equipment and storage medium
CN114764732A (en) * 2020-12-30 2022-07-19 美的集团股份有限公司 Commodity recommendation method and equipment and computer-readable storage medium
CN114781348A (en) * 2022-04-11 2022-07-22 中国电子科技集团公司第十研究所 Text similarity calculation method and system based on bag-of-words model

Patent Citations (20)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120066072A1 (en) * 2007-05-25 2012-03-15 Issar Amit Kanigsberg Recommendation Systems and Methods Using Interest Correlation
US20110307813A1 (en) * 2010-06-11 2011-12-15 International Business Machines Corporation Interactive Ring-Shaped Interface
US20140101243A1 (en) * 2012-10-05 2014-04-10 Facebook, Inc. Method and apparatus for identifying common interest between social network users
CN103235824A (en) * 2013-05-06 2013-08-07 上海河广信息科技有限公司 Method and system for determining web page texts users interested in according to browsed web pages
CN103235823A (en) * 2013-05-06 2013-08-07 上海河广信息科技有限公司 Method and system for determining current interest of users according to related web pages and current behaviors
US20150120722A1 (en) * 2013-10-31 2015-04-30 Telefonica Digital Espana, S.L.U. Method and system for providing multimedia content recommendations
CN105142028A (en) * 2015-07-29 2015-12-09 华中科技大学 Television program content searching and recommending method oriented to integration of three networks
CN105677846A (en) * 2016-01-06 2016-06-15 中国传媒大学 Recommendation system and construction method thereof
US10394885B1 (en) * 2016-03-15 2019-08-27 Intuit Inc. Methods, systems and computer program products for generating personalized financial podcasts
WO2017193323A1 (en) * 2016-05-12 2017-11-16 深圳大学 User preference-based personalized recommendation method and system utilizing same
US20180197090A1 (en) * 2017-01-10 2018-07-12 Facebook, Inc. Externally informed counterfactual prediction
CN110362740A (en) * 2019-06-10 2019-10-22 河海大学 A kind of water conservancy gateway information mixed recommendation method
CN110263106A (en) * 2019-06-25 2019-09-20 中国人民解放军国防科技大学 Collaborative public opinion fraud detection method and device
WO2021070681A1 (en) * 2019-10-10 2021-04-15 株式会社村田製作所 Degree-of-interest evaluation system and degree-of-interest evaluation method
CN111079028A (en) * 2019-12-04 2020-04-28 上海财经大学 Collaborative filtering recommendation system and method based on multi-source auxiliary information
CN112269932A (en) * 2020-10-30 2021-01-26 金天国际医疗科技有限公司 Big data-based small and medium enterprise resource integration processing system
CN114764732A (en) * 2020-12-30 2022-07-19 美的集团股份有限公司 Commodity recommendation method and equipment and computer-readable storage medium
CN112632397A (en) * 2021-01-04 2021-04-09 同方知网(北京)技术有限公司 Personalized recommendation method based on multi-type academic achievement portrait and mixed recommendation strategy
CN113988060A (en) * 2021-08-31 2022-01-28 中国电子科技集团公司第十五研究所 Data information recommendation method and device, terminal equipment and storage medium
CN114781348A (en) * 2022-04-11 2022-07-22 中国电子科技集团公司第十研究所 Text similarity calculation method and system based on bag-of-words model

Non-Patent Citations (6)

* Cited by examiner, † Cited by third party
Title
CHUNLIANG LU 等: "Twitter User Modeling and Tweets Recommendation Based on Wikipedia Concept Graph", 《WORKSHOPS AT THE TWENTY-SIXTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE》 *
JINGHUAZHAO 等: "User personality prediction based on topic preference and sentiment analysis using LSTM model", 《PATTERN RECOGNITION LETTERS》 *
张培颖: "基于模糊隶属度的个性化网页推荐***", 《计算机***应用》 *
徐麟军 等: "学习资源个性化检索***中用户兴趣建模研究", 《电脑知识与技术》 *
葛桂丽 等: "基于情境感知的用户个性化兴趣建模", 《计算机应用研究》 *
陆洲 等: "基于用户兴趣模型聚类的协同过滤推荐算法", 《微计算机信息》 *

Also Published As

Publication number Publication date
CN114971817B (en) 2022-11-22

Similar Documents

Publication Publication Date Title
US11748397B2 (en) Dynamic feedback in a recommendation system
JP4650541B2 (en) RECOMMENDATION DEVICE AND METHOD, PROGRAM, AND RECORDING MEDIUM
JP4678546B2 (en) RECOMMENDATION DEVICE AND METHOD, PROGRAM, AND RECORDING MEDIUM
JP4538757B2 (en) Information processing apparatus, information processing method, and program
US20150242750A1 (en) Asymmetric Rankers for Vector-Based Recommendation
WO2016077255A1 (en) Hybrid explanations in collaborative filter based recommendation system
CN112328849A (en) User portrait construction method, user portrait-based dialogue method and device
CN112950325B (en) Self-attention sequence recommendation method for social behavior fusion
CN111858969B (en) Multimedia data recommendation method, device, computer equipment and storage medium
CN111488526A (en) Recommendation method and device
Su et al. Effective social content-based collaborative filtering for music recommendation
CN115455280A (en) Recommendation list determining method and server
CN115618024A (en) Multimedia recommendation method and device and electronic equipment
JP2024508502A (en) Methods and devices for pushing information
CN107506479A (en) A kind of object recommendation method and apparatus
JP5013161B2 (en) Information processing apparatus, information processing method, providing apparatus, providing method, and program
CN110569447B (en) Network resource recommendation method and device and storage medium
CN113836390A (en) Resource recommendation method and device, computer equipment and storage medium
CN114971817B (en) Product self-adaptive service method, medium and device based on user demand portrait
El Alami et al. Improving Neighborhood-Based Collaborative Filtering by a Heuristic Approach and an Adjusted Similarity Measure.
CN112035740A (en) Project use duration prediction method, device, equipment and storage medium
JP5625792B2 (en) Information processing apparatus, latent feature amount calculation method, and program
CN111563159B (en) Text sorting method and device
CN117909542A (en) Video recommendation method, device, equipment and storage medium
CN117093774A (en) Cross-domain recommendation and model training method, device, 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
GR01 Patent grant
GR01 Patent grant