CN113704608A - Personalized item recommendation method and device, electronic equipment and storage medium - Google Patents
Personalized item recommendation method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN113704608A CN113704608A CN202110991007.4A CN202110991007A CN113704608A CN 113704608 A CN113704608 A CN 113704608A CN 202110991007 A CN202110991007 A CN 202110991007A CN 113704608 A CN113704608 A CN 113704608A
- Authority
- CN
- China
- Prior art keywords
- project
- user
- target
- matrix
- determining
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 40
- 239000011159 matrix material Substances 0.000 claims abstract description 138
- 238000012545 processing Methods 0.000 claims description 19
- 238000004422 calculation algorithm Methods 0.000 claims description 17
- 238000004590 computer program Methods 0.000 claims description 11
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims description 6
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 11
- 238000013507 mapping Methods 0.000 description 10
- 238000011156 evaluation Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 4
- 238000005192 partition Methods 0.000 description 4
- 238000004891 communication Methods 0.000 description 3
- 238000000388 lateral force microscopy Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 239000003086 colorant Substances 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000007599 discharging Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Images
Classifications
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Probability & Statistics with Applications (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention relates to the field of artificial intelligence, and provides a personalized project recommendation method, a personalized project recommendation device, electronic equipment and a storage medium, wherein the method comprises the following steps: acquiring user data corresponding to a target user and a plurality of associated users, and constructing a project scoring matrix and a project display factor matrix based on the user data; inputting the project scoring matrix and the project explicit factor matrix into a trained implicit semantic model to obtain a project implicit factor matrix; splicing the project explicit factor matrix and the project implicit factor matrix to obtain a project attribute matrix; carrying out fuzzy clustering on the project attribute matrix to obtain a clustering result; and determining the score similarity between the target user and each associated user according to the clustering result, and generating the item recommendation information corresponding to the target user based on the score similarity. The method and the device can improve the accuracy of project recommendation.
Description
Technical Field
The invention relates to the field of artificial intelligence, in particular to a personalized project recommendation method and device, electronic equipment and a storage medium.
Background
With the development of the internet, the convenience of various information platforms such as websites and applications enables users to release and acquire items such as movie items, game items, music items, short video items and the like anytime and anywhere, which also causes the information platforms such as websites and applications to present an overloaded state of items. A large number of items are really valuable wealth for various information platforms, but meanwhile, the difficulty of the information platforms in recommending personalized items to users is increased due to the overload of the items, and the users are inundated with the items in different shapes and colors when using the information platforms, so that the real interested items are probably missed.
Therefore, how to depersonalize and recommend items meeting the user interests for each user of the information platform is a problem worthy of research.
Disclosure of Invention
In view of the above, there is a need for a method, an apparatus, an electronic device and a storage medium for personalized item recommendation, which can improve the accuracy of item recommendation.
A first aspect of the present invention provides a method for personalized item recommendation, the method comprising:
acquiring user data corresponding to a target user and a plurality of associated users, and constructing a project scoring matrix and a project display factor matrix based on the user data;
inputting the project scoring matrix and the project explicit factor matrix into a trained implicit semantic model to obtain a project implicit factor matrix;
splicing the project explicit factor matrix and the project implicit factor matrix to obtain a project attribute matrix;
carrying out fuzzy clustering on the project attribute matrix to obtain a clustering result;
and determining the score similarity between the target user and each associated user according to the clustering result, and generating the item recommendation information corresponding to the target user based on the score similarity.
According to the technical scheme, user data corresponding to a target user and a plurality of associated users are obtained, and a project scoring matrix and a project display factor matrix are constructed on the basis of the user data; inputting the project scoring matrix and the project explicit factor matrix into a trained latent semantic model to obtain a project implicit factor matrix, wherein the project implicit factor matrix comprises hidden preferences corresponding to a target user and a plurality of associated users; then, the project explicit factor matrix and the project implicit factor matrix are spliced to obtain a project attribute matrix, so that the project attribute matrix obtained through splicing contains more user information, and the recommendation accuracy is improved; then carrying out fuzzy clustering on the project attribute matrix to obtain a clustering result; and finally, according to the clustering result, determining the score similarity between the target user and each associated user, and generating the project recommendation information corresponding to the target user based on the score similarity, so that projects meeting the user interests can be recommended to the target user in a personalized manner, and the project recommendation accuracy is improved.
According to an optional embodiment of the present invention, the performing fuzzy clustering on the item attribute matrix to obtain a clustering result includes:
determining an initial clustering center according to the user data;
and carrying out fuzzy clustering on the project attribute matrix based on the initial clustering center to obtain a clustering result.
According to the technical scheme, the initial clustering center is determined through the user data, the accuracy of determining the initial clustering center is improved, and the situation that the initial clustering center is improperly selected to cause that the initial clustering center is trapped in a local extreme point when fuzzy clustering is carried out and a global optimal solution cannot be obtained is avoided. Meanwhile, the situations that the iteration times are increased, the time complexity is improved and the recommendation efficiency is reduced after the initialization clustering center is randomly selected due to data sparseness in the project attribute matrix are avoided.
According to an optional embodiment of the present invention, the determining an initial cluster center according to the user data comprises:
determining a data dimension corresponding to the user data;
carrying out equidistant division processing on the user data according to the data dimension to obtain a plurality of grid objects;
calculating the grid density corresponding to each grid object, and determining a target grid object in the grid objects based on the grid density;
calculating an associated grid object associated with the target grid object based on a breadth-first traversal algorithm, and obtaining a cluster according to the target grid object and the associated grid object;
and determining the center of the cluster as an initial cluster center.
According to the technical scheme, the accuracy of determining the initial clustering center can be effectively improved through the network clustering and breadth-first traversal algorithm, and meanwhile, the efficiency of determining the initial clustering center can also be improved.
According to an optional embodiment of the present invention, the generating of the item recommendation information corresponding to the target user based on the score similarity includes:
determining a plurality of target associated users in the plurality of associated users according to the degree of the grading similarity;
determining a project corresponding to each target associated user, and calculating the interest degree of each target associated user and the corresponding project;
and determining an item to be recommended in the items based on the interest degree, and generating item recommendation information corresponding to the target user according to the item to be recommended.
According to the technical scheme, the associated user with high similarity is determined to be the target associated user in the plurality of associated users through grading the similarity, the item recommendation information is generated based on the item with high interest degree of the target associated user, and the item recommendation efficiency can be improved.
According to an optional embodiment of the present invention, each target associated user includes a plurality of items, and the calculating the interest level of each target associated user and the corresponding item includes:
calculating the Euclidean distance between each target associated user and each item;
and determining the interest degree of the target associated user and the item according to the Euclidean distance.
According to the technical scheme, the interest degree of the target associated user and the item is determined according to the Euclidean distance between the target associated user and the item, and the efficiency of determining the interest degree is improved.
According to an optional embodiment of the present invention, the determining, according to the clustering result, the score similarity between the target user and each associated user includes:
and determining the scoring similarity between the target user and each associated user by using a cosine similarity algorithm according to the clustering result.
According to the technical scheme, the accuracy rate of calculating the score similarity is improved by using a cosine similarity calculation method.
According to an alternative embodiment of the invention, the method further comprises:
acquiring a plurality of dimension information corresponding to a target user based on a user identifier of the target user;
analyzing the dimension information to determine the association relationship among the dimension information;
constructing a target user portrait corresponding to the target user based on the incidence relation;
determining a plurality of associated users based on the target user representation.
According to the technical scheme, the constructed target user image keeps more user information through the association relation among the plurality of dimensional information, so that the accuracy of determining the associated user is improved.
A second aspect of the present invention provides a personalized item recommendation apparatus, the apparatus comprising:
the system comprises a construction module, a project display factor calculation module and a project display factor calculation module, wherein the construction module is used for acquiring user data corresponding to a target user and a plurality of associated users, and constructing a project scoring matrix and a project display factor matrix based on the user data;
the processing module is used for inputting the project scoring matrix and the project explicit factor matrix into a trained implicit semantic model to obtain a project implicit factor matrix;
the splicing module is used for splicing the project explicit factor matrix and the project implicit factor matrix to obtain a project attribute matrix;
the clustering module is used for carrying out fuzzy clustering on the project attribute matrix to obtain a clustering result;
and the recommending module is used for determining the scoring similarity between the target user and each associated user according to the clustering result and generating the item recommending information corresponding to the target user based on the scoring similarity.
A third aspect of the invention provides an electronic device comprising a processor and a memory, the processor being adapted to implement the personalized item recommendation method when executing a computer program stored in the memory.
A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the personalized item recommendation method.
In addition, for technical effects brought by the second aspect, the third aspect and the fourth aspect, reference may be made to the description related to the methods designed in the above method part, and details are not repeated here.
Drawings
Fig. 1 is a flowchart of a personalized item recommendation method according to an embodiment of the present invention.
Fig. 2 is a block diagram of a personalized item recommendation apparatus according to a second embodiment of the present invention.
Fig. 3 is a schematic structural diagram of an electronic device according to a third embodiment of the present invention.
Detailed Description
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, not all, embodiments of the present 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.
The flow diagrams depicted in the figures are merely illustrative and do not necessarily include all of the elements and operations/steps, nor do they necessarily have to be performed in the order depicted. For example, some operations/steps may be decomposed, combined or partially combined, so that the actual execution sequence may be changed according to the actual situation.
The embodiment of the invention provides a personalized item recommendation method and device, electronic equipment and a computer readable storage medium. The personalized project recommendation method can be applied to terminal equipment or a server, the terminal equipment can be electronic equipment such as a mobile phone, a tablet personal computer, a notebook computer, a desktop computer, a personal digital assistant and wearable equipment, and the server can be a single server or a server cluster consisting of a plurality of servers. The following explains the personalized item recommendation method applied to a server as an example.
Some embodiments of the invention are described in detail below with reference to the accompanying drawings. The embodiments described below and the features of the embodiments can be combined with each other without conflict.
Referring to fig. 1, fig. 1 is a schematic flow chart of a personalized item recommendation method according to an embodiment of the present invention.
As shown in fig. 1, the personalized item recommendation method specifically includes steps S11 to S15, and the order of the steps in the flowchart may be changed or some of the steps may be omitted according to different requirements.
S11, user data corresponding to the target user and the multiple associated users are obtained, and a project scoring matrix and a project display factor matrix are constructed based on the user data.
The target user is a user to be subjected to item recommendation. The associated user is a user associated with the target user, wherein the association may include a preference association, a platform association, and an identity association. For example, if a user is associated with the target user's preferences, the user is determined to be the target user; if one user and the target user are colleagues, relatives or classmates, determining that the user is the target user; and if one user and the target user are in the same social platform, determining that the user is the target user.
The user data may include user viewing data, user reading data, user entertainment data. The user data includes a plurality of items, and scores for the plurality of items. The project scoring matrix comprises the mapping relation between projects and scores, and the project factor display matrix comprises the mapping relation between the projects and the factor displays. The apparent factor is used for representing the obvious favorite features of the user, and the apparent factor is a factor directly influencing the evaluation of the user on the information. Given that some features are of a certain type, which the user likes or which the item contains, these features may be referred to as explicit factors, since the features are explicit. For example, movies include features such as horror, comedy, love, laugh, action, etc., which are explicit and are therefore referred to as explicit factors.
In an optional embodiment, the method further comprises:
acquiring a plurality of dimension information corresponding to a target user based on a user identifier of the target user;
analyzing the dimension information to determine the association relationship among the dimension information;
constructing a target user portrait corresponding to the target user based on the incidence relation;
determining a plurality of associated users based on the target user representation.
The user identifier may include an identity identifier, a number identifier, a platform identifier, and the like, and for example, data information corresponding to a target user may be obtained based on a name of the user to obtain a plurality of dimensional information; the data information corresponding to the target user can be acquired based on the mobile phone number of the user, and a plurality of dimension information is acquired; and also acquiring data information of the user on the social platform based on the platform identification of the user to obtain a plurality of dimension information. The dimension information is used for presenting the preference of the target user, and can comprise a film viewing dimension, a reading dimension, a game dimension, a motion dimension and the like.
And calculating the association relation between each pair of dimension information, and constructing the target user portrait corresponding to the target user based on the association relation. And constructing the target user image according to the strength of the association relation. For example, several pieces of dimensional information with strong association relations may be used to construct a target user image corresponding to a target user.
Through the incidence relation among the dimensional information, the constructed target user image keeps more user information, and therefore the accuracy of determining the incidence user is improved.
And S12, inputting the project scoring matrix and the project explicit factor matrix into a trained implicit semantic model to obtain a project implicit factor matrix.
Hidden semantic models, also known as hidden factor models (LFMs), are used to find potential topics or classifications in input material, and hidden factors can be determined. Hidden factors are factors that potentially affect the evaluation of information by a user. For example, each user is considered to have own preference information, and each information also contains preference information of all users, so that the high scoring of the information by the user is embodied that the preference information contained in the information is just the preference information of the user, but the preference information cannot be determined definitely, and therefore, the preference information can be determined as a factor which potentially affects the evaluation of the information by the user, namely a hidden factor.
The project implicit factor matrix comprises mapping relations between projects and implicit factors.
And S13, splicing the project explicit factor matrix and the project implicit factor matrix to obtain a project attribute matrix.
And performing matrix splicing on the project explicit factor matrix and the project implicit factor matrix to obtain a project attribute matrix, wherein the project attribute matrix comprises a mapping relation between the projects and the explicit factors and a mapping relation between the projects and the implicit factors.
And S14, carrying out fuzzy clustering on the project attribute matrix to obtain a clustering result.
Fuzzy clustering processing can be carried out on the project attribute matrix based on a fuzzy c-means clustering algorithm (FCM), the membership degree of each factor in the project attribute matrix to all class centers is obtained by optimizing an objective function, and the factors in the project attribute matrix are automatically classified to obtain a clustering result.
In an optional embodiment, the performing fuzzy clustering on the item attribute matrix to obtain a clustering result includes:
determining an initial clustering center according to the user data;
and carrying out fuzzy clustering on the project attribute matrix based on the initial clustering center to obtain a clustering result.
A data type corresponding to the user data may be determined, and an initial cluster center may be determined based on the data type.
The initial clustering center is determined through the user data, the accuracy of determining the initial clustering center is improved, and the situation that the initial clustering center is improperly selected to cause that the initial clustering center falls into a local extreme point and the global optimal solution cannot be obtained is avoided. Meanwhile, the situations that the iteration times are increased, the time complexity is improved and the recommendation efficiency is reduced after the initialization clustering center is randomly selected due to data sparseness in the project attribute matrix are avoided.
In an optional embodiment, the determining an initial cluster center according to the user data includes:
determining a data dimension corresponding to the user data;
carrying out equidistant division processing on the user data according to the data dimension to obtain a plurality of grid objects;
calculating the grid density corresponding to each grid object, and determining a target grid object in the grid objects based on the grid density;
calculating an associated grid object associated with the target grid object based on a breadth-first traversal algorithm, and obtaining a cluster according to the target grid object and the associated grid object;
and determining the center of the cluster as an initial cluster center.
The data dimension corresponding to the user may be determined based on the data type corresponding to the user data. For example, the user data types may include movies and their corresponding ratings, games and their corresponding ratings, images and their corresponding ratings. The same data type is the same dimension. The number of data types matches the data dimensions, with each data type corresponding to a data dimension.
Illustratively, if a plurality of data dimensions exist in user data, the equidistantly dividing the user data according to the data dimensions to obtain a plurality of grid objects includes: generating a plurality of grid cells according to the plurality of data dimensions, wherein each data dimension has one and only one corresponding grid cell; and according to the data characteristics corresponding to each data dimension, carrying out equidistant partition processing on the grid cells corresponding to the data dimension, and filling the user data corresponding to the data dimension into the grid cells subjected to the equidistant partition processing to obtain a plurality of grid objects. The grid object may be a cell that includes user data.
The division distances in the equidistant division process for different data features are also different. For example, the user data includes three data dimensions, and three grid cells can be obtained by performing equidistant division processing on the user data according to the three data dimensions.
The mesh density is a ratio of the number of user data to the total number of user data in each mesh object. User data at the edges of a grid object may be considered user data in the grid object. For example, the mesh object with the highest mesh density in each mesh cell may be determined as the target mesh object.
And traversing the grid objects around the target grid object based on a breadth-first traversal algorithm, and determining an associated grid object associated with the target grid object. And taking the associated grid object and the target grid object obtained by traversing as a cluster, namely determining all connected networks obtained by traversing as a cluster. And determining the center of the cluster as an initial cluster center.
If a plurality of data dimensions exist in the user data, a plurality of clustering clusters can be obtained, and a plurality of initial clustering centers can be obtained.
The accuracy of determining the initial clustering center can be effectively improved through the network clustering and breadth-first traversal algorithm, and meanwhile, the efficiency of determining the initial clustering center can also be improved.
And S15, determining the score similarity of the target user and each associated user according to the clustering result, and generating the item recommendation information corresponding to the target user based on the score similarity.
Illustratively, according to the clustering result, the scoring similarity between the target user and each associated user is determined by using a cosine similarity algorithm. The accuracy rate of calculating the score similarity is improved by using the cosine similarity algorithm.
In an optional embodiment, the generating of the item recommendation information corresponding to the target user based on the score similarity includes:
determining a plurality of target associated users in the plurality of associated users according to the degree of the grading similarity;
determining a project corresponding to each target associated user, and calculating the interest degree of each target associated user and the corresponding project;
and determining an item to be recommended in the items based on the interest degree, and generating item recommendation information corresponding to the target user according to the item to be recommended.
Illustratively, the multiple associated users are ranked according to the degree of the score similarity, and multiple target associated users are determined in the multiple associated users according to a ranking result. For example, the first Z associated users are determined as target associated users according to the sorting result.
And determining the item corresponding to each target associated user according to the user data. Illustratively, the items to be recommended are determined in the items corresponding to the target associated user according to the level of the interest degree. For example, the item with high user interest degree in the item corresponding to the target associated user is determined as the item to be recommended.
And determining the associated user with high similarity as the target associated user in the plurality of associated users by scoring the similarity, and generating the item recommendation information based on the item with high interest degree of the target associated user, so that the item recommendation efficiency can be improved.
In an optional embodiment, each target associated user includes a plurality of items, and the calculating the interest level of each target associated user and the corresponding item includes:
calculating the Euclidean distance between each target associated user and each item;
and determining the interest degree of the target associated user and the item according to the Euclidean distance.
The smaller the Euclidean distance is, the higher the interest degree of the target associated user and the item is determined to be; the greater the Euclidean distance, the lower the level of interest in determining the target associated user with the item.
The interest degree of the target associated user and the item is determined through the Euclidean distance between the target associated user and the item, and the efficiency of determining the interest degree is improved.
According to the personalized project recommendation method provided by the embodiment, user data corresponding to a target user and a plurality of associated users are obtained, and a project scoring matrix and a project display factor matrix are constructed based on the user data; inputting the project scoring matrix and the project explicit factor matrix into a trained latent semantic model to obtain a project implicit factor matrix, wherein the project implicit factor matrix comprises hidden preferences corresponding to a target user and a plurality of associated users; then, the project explicit factor matrix and the project implicit factor matrix are spliced to obtain a project attribute matrix, so that the project attribute matrix obtained through splicing contains more user information, and the recommendation accuracy is improved; then carrying out fuzzy clustering on the project attribute matrix to obtain a clustering result; and finally, according to the clustering result, determining the score similarity between the target user and each associated user, and generating the project recommendation information corresponding to the target user based on the score similarity, so that projects meeting the user interests can be recommended to the target user in a personalized manner, and the project recommendation accuracy is improved.
Referring to fig. 2, fig. 2 is a schematic block diagram of a personalized item recommendation apparatus according to an embodiment of the present invention.
In some embodiments, the personalized item recommendation device 20 may include a plurality of functional modules consisting of computer program segments, the computer program of each program segment in the personalized item recommendation device 20 may be stored in a memory of the electronic device and executed by at least one processor to perform (see fig. 1 for details) the function of personalized item recommendation, and the personalized item recommendation device may be configured in a server or a terminal.
The server may be an independent server or a server cluster. The terminal can be an electronic device such as a mobile phone, a tablet computer, a notebook computer, a desktop computer, a personal digital assistant and a wearable device.
In this embodiment, the personalized item recommendation device 20 may be divided into a plurality of functional modules according to the functions it performs. As shown in fig. 2, the functional module may include: a construction module 201, a processing module 202, a concatenation module 203, a clustering module 204, and a recommendation module 205. The module referred to herein is a series of computer program segments capable of being executed by at least one processor and capable of performing a fixed function and is stored in memory. In the present embodiment, the functions of the modules will be described in detail in the following embodiments.
The construction module 201 is configured to obtain user data corresponding to a target user and multiple associated users, and construct a project scoring matrix and a project display factor matrix based on the user data.
The target user is a user to be subjected to item recommendation. The associated user is a user associated with the target user, wherein the association may include a preference association, a platform association, and an identity association. For example, if a user is associated with the target user's preferences, the user is determined to be the target user; if one user and the target user are colleagues, relatives or classmates, determining that the user is the target user; and if one user and the target user are in the same social platform, determining that the user is the target user.
The user data may include user viewing data, user reading data, user entertainment data. The user data includes a plurality of items, and scores for the plurality of items. The project scoring matrix comprises the mapping relation between projects and scores, and the project factor display matrix comprises the mapping relation between the projects and the factor displays. The apparent factor is used for representing the obvious favorite features of the user, and the apparent factor is a factor directly influencing the evaluation of the user on the information. Given that some features are of a certain type, which the user likes or which the item contains, these features may be referred to as explicit factors, since the features are explicit. For example, movies include features such as horror, comedy, love, laugh, action, etc., which are explicit and are therefore referred to as explicit factors.
In an alternative embodiment, the building module 201 is further configured to:
acquiring a plurality of dimension information corresponding to a target user based on a user identifier of the target user;
analyzing the dimension information to determine the association relationship among the dimension information;
constructing a target user portrait corresponding to the target user based on the incidence relation;
determining a plurality of associated users based on the target user representation.
The user identifier may include an identity identifier, a number identifier, a platform identifier, and the like, and for example, data information corresponding to a target user may be obtained based on a name of the user to obtain a plurality of dimensional information; the data information corresponding to the target user can be acquired based on the mobile phone number of the user, and a plurality of dimension information is acquired; and also acquiring data information of the user on the social platform based on the platform identification of the user to obtain a plurality of dimension information. The dimension information is used for presenting the preference of the target user, and can comprise a film viewing dimension, a reading dimension, a game dimension, a motion dimension and the like.
And calculating the association relation between each pair of dimension information, and constructing the target user portrait corresponding to the target user based on the association relation. And constructing the target user image according to the strength of the association relation. For example, several pieces of dimensional information with strong association relations may be used to construct a target user image corresponding to a target user.
Through the incidence relation among the dimensional information, the constructed target user image keeps more user information, and therefore the accuracy of determining the incidence user is improved.
And the processing module 202 is configured to input the project scoring matrix and the project explicit factor matrix into a trained implicit semantic model to obtain a project implicit factor matrix.
Hidden semantic models, also known as hidden factor models (LFMs), are used to find potential topics or classifications in input material, and hidden factors can be determined. Hidden factors are factors that potentially affect the evaluation of information by a user. For example, each user is considered to have own preference information, and each information also contains preference information of all users, so that the high scoring of the information by the user is embodied that the preference information contained in the information is just the preference information of the user, but the preference information cannot be determined definitely, and therefore, the preference information can be determined as a factor which potentially affects the evaluation of the information by the user, namely a hidden factor.
The project implicit factor matrix comprises mapping relations between projects and implicit factors.
And the splicing module 203 is used for splicing the project explicit factor matrix and the project implicit factor matrix to obtain a project attribute matrix.
And performing matrix splicing on the project explicit factor matrix and the project implicit factor matrix to obtain a project attribute matrix, wherein the project attribute matrix comprises a mapping relation between the projects and the explicit factors and a mapping relation between the projects and the implicit factors.
And the clustering module 204 is used for performing fuzzy clustering on the item attribute matrix to obtain a clustering result.
Fuzzy clustering processing can be carried out on the project attribute matrix based on a fuzzy c-means clustering algorithm (FCM), the membership degree of each factor in the project attribute matrix to all class centers is obtained by optimizing an objective function, and the factors in the project attribute matrix are automatically classified to obtain a clustering result.
In an optional embodiment, the clustering module 204 performs fuzzy clustering on the item attribute matrix, and obtaining a clustering result includes:
determining an initial clustering center according to the user data;
and carrying out fuzzy clustering on the project attribute matrix based on the initial clustering center to obtain a clustering result.
A data type corresponding to the user data may be determined, and an initial cluster center may be determined based on the data type.
The initial clustering center is determined through the user data, the accuracy of determining the initial clustering center is improved, and the situation that the initial clustering center is improperly selected to cause that the initial clustering center falls into a local extreme point and the global optimal solution cannot be obtained is avoided. Meanwhile, the situations that the iteration times are increased, the time complexity is improved and the recommendation efficiency is reduced after the initialization clustering center is randomly selected due to data sparseness in the project attribute matrix are avoided.
In an alternative embodiment, the clustering module 204 determines the initial cluster center according to the user data, including:
determining a data dimension corresponding to the user data;
carrying out equidistant division processing on the user data according to the data dimension to obtain a plurality of grid objects;
calculating the grid density corresponding to each grid object, and determining a target grid object in the grid objects based on the grid density;
calculating an associated grid object associated with the target grid object based on a breadth-first traversal algorithm, and obtaining a cluster according to the target grid object and the associated grid object;
and determining the center of the cluster as an initial cluster center.
The data dimension corresponding to the user may be determined based on the data type corresponding to the user data. For example, the user data types may include movies and their corresponding ratings, games and their corresponding ratings, images and their corresponding ratings. The same data type is the same dimension. The number of data types matches the data dimensions, with each data type corresponding to a data dimension.
Illustratively, if a plurality of data dimensions exist in user data, the equidistantly dividing the user data according to the data dimensions to obtain a plurality of grid objects includes: generating a plurality of grid cells according to the plurality of data dimensions, wherein each data dimension has one and only one corresponding grid cell; and according to the data characteristics corresponding to each data dimension, carrying out equidistant partition processing on the grid cells corresponding to the data dimension, and filling the user data corresponding to the data dimension into the grid cells subjected to the equidistant partition processing to obtain a plurality of grid objects. The grid object may be a cell that includes user data.
The division distances in the equidistant division process for different data features are also different. For example, the user data includes three data dimensions, and three grid cells can be obtained by performing equidistant division processing on the user data according to the three data dimensions.
The mesh density is a ratio of the number of user data to the total number of user data in each mesh object. User data at the edges of a grid object may be considered user data in the grid object. For example, the mesh object with the highest mesh density in each mesh cell may be determined as the target mesh object.
And traversing the grid objects around the target grid object based on a breadth-first traversal algorithm, and determining an associated grid object associated with the target grid object. And taking the associated grid object and the target grid object obtained by traversing as a cluster, namely determining all connected networks obtained by traversing as a cluster. And determining the center of the cluster as an initial cluster center.
If a plurality of data dimensions exist in the user data, a plurality of clustering clusters can be obtained, and a plurality of initial clustering centers can be obtained.
The accuracy of determining the initial clustering center can be effectively improved through the network clustering and breadth-first traversal algorithm, and meanwhile, the efficiency of determining the initial clustering center can also be improved.
And the recommending module 205 is configured to determine the score similarity between the target user and each associated user according to the clustering result, and generate item recommendation information corresponding to the target user based on the score similarity.
Illustratively, according to the clustering result, the scoring similarity between the target user and each associated user is determined by using a cosine similarity algorithm. The accuracy rate of calculating the score similarity is improved by using the cosine similarity algorithm.
In an optional embodiment, the generating, by the recommendation module 205, the item recommendation information corresponding to the target user based on the score similarity includes:
determining a plurality of target associated users in the plurality of associated users according to the degree of the grading similarity;
determining a project corresponding to each target associated user, and calculating the interest degree of each target associated user and the corresponding project;
and determining an item to be recommended in the items based on the interest degree, and generating item recommendation information corresponding to the target user according to the item to be recommended.
Illustratively, the multiple associated users are ranked according to the degree of the score similarity, and multiple target associated users are determined in the multiple associated users according to a ranking result. For example, the first Z associated users are determined as target associated users according to the sorting result.
And determining the item corresponding to each target associated user according to the user data. Illustratively, the items to be recommended are determined in the items corresponding to the target associated user according to the level of the interest degree. For example, the item with high user interest degree in the item corresponding to the target associated user is determined as the item to be recommended.
And determining the associated user with high similarity as the target associated user in the plurality of associated users by scoring the similarity, and generating the item recommendation information based on the item with high interest degree of the target associated user, so that the item recommendation efficiency can be improved.
In an alternative embodiment, each target associated user includes a plurality of items, and the calculating, by the recommending module 205, the interest level of each target associated user in the corresponding item includes:
calculating the Euclidean distance between each target associated user and each item;
and determining the interest degree of the target associated user and the item according to the Euclidean distance.
The smaller the Euclidean distance is, the higher the interest degree of the target associated user and the item is determined to be; the greater the Euclidean distance, the lower the level of interest in determining the target associated user with the item.
The interest degree of the target associated user and the item is determined through the Euclidean distance between the target associated user and the item, and the efficiency of determining the interest degree is improved.
The personalized item recommendation apparatus provided in the above embodiments may be implemented in the form of a computer program, which can be run on the electronic device as shown in fig. 3. The electronic device 3 comprises a memory 31, at least one processor 32, at least one communication bus 33 and a transceiver 34.
It will be appreciated by those skilled in the art that the configuration of the electronic device shown in fig. 3 does not constitute a limitation of the embodiments of the present application, and may be a bus-type configuration or a star-type configuration, and that the electronic device 3 may include more or less hardware or software than those shown, or a different arrangement of components.
In some embodiments, the electronic device 3 is a device capable of automatically performing numerical calculation and/or information processing according to instructions set or stored in advance, and the hardware thereof includes but is not limited to a microprocessor, an application specific integrated circuit, a programmable gate array, a digital processor, an embedded device, and the like. The electronic device 3 may also include a client device, which includes, but is not limited to, any electronic product that can interact with a client through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device, for example, a personal computer, a tablet computer, a smart phone, a digital camera, and the like.
It should be noted that the electronic device 3 is only an example, and other existing or future electronic products, such as those that can be adapted to the present application, should also be included in the scope of protection of the present application, and are included by reference.
In some embodiments, the memory 31 has stored therein a computer program that, when executed by the at least one processor 32, performs the steps of:
acquiring user data corresponding to a target user and a plurality of associated users, and constructing a project scoring matrix and a project display factor matrix based on the user data;
inputting the project scoring matrix and the project explicit factor matrix into a trained implicit semantic model to obtain a project implicit factor matrix;
splicing the project explicit factor matrix and the project implicit factor matrix to obtain a project attribute matrix;
carrying out fuzzy clustering on the project attribute matrix to obtain a clustering result;
and determining the score similarity between the target user and each associated user according to the clustering result, and generating the item recommendation information corresponding to the target user based on the score similarity.
The Memory 31 includes a Read-Only Memory (ROM), a Programmable Read-Only Memory (PROM), an Erasable Programmable Read-Only Memory (EPROM), a One-time Programmable Read-Only Memory (OTPROM), an electronically Erasable rewritable Read-Only Memory (Electrically-Erasable Programmable Read-Only Memory (EEPROM)), an optical Read-Only disk (CD-ROM) or other optical disk Memory, a magnetic disk Memory, a tape Memory, or any other medium readable by a computer capable of carrying or storing data.
Further, the computer-readable storage medium may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function, and the like; the storage data area may store data created according to the use of the blockchain node, and the like.
In some embodiments, the at least one processor 32 is a Control Unit (Control Unit) of the electronic device 3, connects various components of the electronic device 3 by various interfaces and lines, and executes various functions and processes data of the electronic device 3 by running or executing programs or modules stored in the memory 31 and calling data stored in the memory 31. For example, the at least one processor 32, when executing the computer program stored in the memory, implements all or a portion of the steps of the personalized item recommendation method described in embodiments of the present invention; or implement all or part of the functionality of the personalized item recommendation device. The at least one processor 32 may be composed of an integrated circuit, for example, a single packaged integrated circuit, or may be composed of a plurality of integrated circuits packaged with the same or different functions, including one or more Central Processing Units (CPUs), microprocessors, digital Processing chips, graphics processors, and combinations of various control chips.
In some embodiments, the at least one communication bus 33 is arranged to enable connection communication between the memory 31 and the at least one processor 32 or the like.
Although not shown, the electronic device 3 may further include a power supply (such as a battery) for supplying power to each component, and preferably, the power supply may be logically connected to the at least one processor 32 through a power management device, so as to implement functions of managing charging, discharging, and power consumption through the power management device. The power supply may also include any component of one or more dc or ac power sources, recharging devices, power failure detection circuitry, power converters or inverters, power status indicators, and the like. The electronic device 3 may further include various sensors, a bluetooth module, a Wi-Fi module, and the like, which are not described herein again.
The present embodiment provides a computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, implements the steps of the above-mentioned personalized item recommendation method embodiment, or which, when being executed by a processor, implements the functions of the modules/units of the above-mentioned apparatus embodiment.
The integrated unit implemented in the form of a software functional module may be stored in a computer-readable storage medium. The software functional module is stored in a storage medium and includes several instructions to enable an electronic device (which may be a personal computer, an electronic device, or a network device) or a processor (processor) to execute parts of the methods according to the embodiments of the present invention.
According to the personalized item recommendation device, the electronic equipment and the computer-readable storage medium provided by the embodiment, the user data corresponding to the target user and the multiple associated users are obtained, and the item scoring matrix and the item display factor matrix are constructed based on the user data; inputting the project scoring matrix and the project explicit factor matrix into a trained latent semantic model to obtain a project implicit factor matrix, wherein the project implicit factor matrix comprises hidden preferences corresponding to a target user and a plurality of associated users; then, the project explicit factor matrix and the project implicit factor matrix are spliced to obtain a project attribute matrix, so that the project attribute matrix obtained through splicing contains more user information, and the recommendation accuracy is improved; then carrying out fuzzy clustering on the project attribute matrix to obtain a clustering result; and finally, according to the clustering result, determining the score similarity between the target user and each associated user, and generating the project recommendation information corresponding to the target user based on the score similarity, so that projects meeting the user interests can be recommended to the target user in a personalized manner, and the project recommendation accuracy is improved.
In the embodiments provided in the present invention, it should be understood that the disclosed apparatus and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the modules is only one logical functional division, and other divisions may be realized in practice.
The modules described as separate parts may or may not be physically separate, and parts displayed as modules may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment.
In addition, functional modules in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional module.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference sign in a claim should not be construed as limiting the claim concerned. Furthermore, it is obvious that the word "comprising" does not exclude other elements or that the singular does not exclude the plural. A plurality of units or means recited in the specification may also be implemented by one unit or means through software or hardware. The terms first, second, etc. are used to denote names, but not any particular order.
Claims (10)
1. A method for personalized item recommendation, the method comprising:
acquiring user data corresponding to a target user and a plurality of associated users, and constructing a project scoring matrix and a project display factor matrix based on the user data;
inputting the project scoring matrix and the project explicit factor matrix into a trained implicit semantic model to obtain a project implicit factor matrix;
splicing the project explicit factor matrix and the project implicit factor matrix to obtain a project attribute matrix;
carrying out fuzzy clustering on the project attribute matrix to obtain a clustering result;
and determining the score similarity between the target user and each associated user according to the clustering result, and generating the item recommendation information corresponding to the target user based on the score similarity.
2. The method of claim 1, wherein the fuzzy clustering of the item attribute matrix to obtain a clustering result comprises:
determining an initial clustering center according to the user data;
and carrying out fuzzy clustering on the project attribute matrix based on the initial clustering center to obtain a clustering result.
3. The personalized item recommendation method of claim 2, wherein the determining an initial cluster center based on the user data comprises:
determining a data dimension corresponding to the user data;
carrying out equidistant division processing on the user data according to the data dimension to obtain a plurality of grid objects;
calculating the grid density corresponding to each grid object, and determining a target grid object in the grid objects based on the grid density;
calculating an associated grid object associated with the target grid object based on a breadth-first traversal algorithm, and obtaining a cluster according to the target grid object and the associated grid object;
and determining the center of the cluster as an initial cluster center.
4. The personalized item recommendation method of claim 1, wherein the generating item recommendation information corresponding to the target user based on the scored similarity comprises:
determining a plurality of target associated users in the plurality of associated users according to the degree of the grading similarity;
determining a project corresponding to each target associated user, and calculating the interest degree of each target associated user and the corresponding project;
and determining an item to be recommended in the items based on the interest degree, and generating item recommendation information corresponding to the target user according to the item to be recommended.
5. The personalized item recommendation method of claim 4, wherein each target associated user comprises a plurality of items, and the calculating of the interest level of each target associated user with its corresponding item comprises:
calculating the Euclidean distance between each target associated user and each item;
and determining the interest degree of the target associated user and the item according to the Euclidean distance.
6. The personalized item recommendation method of claim 1, wherein the determining the score similarity of the target user and each associated user according to the clustering result comprises:
and determining the scoring similarity between the target user and each associated user by using a cosine similarity algorithm according to the clustering result.
7. The personalized item recommendation method of claim 1, the method further comprising:
acquiring a plurality of dimension information corresponding to a target user based on a user identifier of the target user;
analyzing the dimension information to determine the association relationship among the dimension information;
constructing a target user portrait corresponding to the target user based on the incidence relation;
determining a plurality of associated users based on the target user representation.
8. A personalized item recommendation apparatus, the apparatus comprising:
the system comprises a construction module, a project display factor calculation module and a project display factor calculation module, wherein the construction module is used for acquiring user data corresponding to a target user and a plurality of associated users, and constructing a project scoring matrix and a project display factor matrix based on the user data;
the processing module is used for inputting the project scoring matrix and the project explicit factor matrix into a trained implicit semantic model to obtain a project implicit factor matrix;
the splicing module is used for splicing the project explicit factor matrix and the project implicit factor matrix to obtain a project attribute matrix;
the clustering module is used for carrying out fuzzy clustering on the project attribute matrix to obtain a clustering result;
and the recommending module is used for determining the scoring similarity between the target user and each associated user according to the clustering result and generating the item recommending information corresponding to the target user based on the scoring similarity.
9. An electronic device, characterized in that the electronic device comprises a processor and a memory, the processor being configured to implement the personalized item recommendation method according to any one of claims 1 to 7 when executing a computer program stored in the memory.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the personalized item recommendation method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110991007.4A CN113704608A (en) | 2021-08-26 | 2021-08-26 | Personalized item recommendation method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110991007.4A CN113704608A (en) | 2021-08-26 | 2021-08-26 | Personalized item recommendation method and device, electronic equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113704608A true CN113704608A (en) | 2021-11-26 |
Family
ID=78655519
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110991007.4A Pending CN113704608A (en) | 2021-08-26 | 2021-08-26 | Personalized item recommendation method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113704608A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112699302A (en) * | 2021-01-05 | 2021-04-23 | 汕头大学 | Real-time recommendation method based on geodesic distance |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104156436A (en) * | 2014-08-13 | 2014-11-19 | 福州大学 | Social association cloud media collaborative filtering and recommending method |
| KR20170011933A (en) * | 2015-07-22 | 2017-02-02 | 주식회사 베티 | System for recommending online lecture |
| US20170206551A1 (en) * | 2016-01-15 | 2017-07-20 | Adobe Systems Incorporated | Personalized Recommendation Computation in Real Time using Incremental Matrix Factorization and User Factor Clustering |
| CN108537274A (en) * | 2018-04-08 | 2018-09-14 | 武汉大学 | A kind of Multi scale Fast Speed Clustering based on grid |
| CN112100512A (en) * | 2020-04-10 | 2020-12-18 | 南京邮电大学 | Collaborative filtering recommendation method based on user clustering and project association analysis |
| CN112955883A (en) * | 2018-12-29 | 2021-06-11 | 深圳市欢太科技有限公司 | Application recommendation method and device, server and computer-readable storage medium |
-
2021
- 2021-08-26 CN CN202110991007.4A patent/CN113704608A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104156436A (en) * | 2014-08-13 | 2014-11-19 | 福州大学 | Social association cloud media collaborative filtering and recommending method |
| KR20170011933A (en) * | 2015-07-22 | 2017-02-02 | 주식회사 베티 | System for recommending online lecture |
| US20170206551A1 (en) * | 2016-01-15 | 2017-07-20 | Adobe Systems Incorporated | Personalized Recommendation Computation in Real Time using Incremental Matrix Factorization and User Factor Clustering |
| CN108537274A (en) * | 2018-04-08 | 2018-09-14 | 武汉大学 | A kind of Multi scale Fast Speed Clustering based on grid |
| CN112955883A (en) * | 2018-12-29 | 2021-06-11 | 深圳市欢太科技有限公司 | Application recommendation method and device, server and computer-readable storage medium |
| CN112100512A (en) * | 2020-04-10 | 2020-12-18 | 南京邮电大学 | Collaborative filtering recommendation method based on user clustering and project association analysis |
Non-Patent Citations (1)
| Title |
|---|
| 杨游云，周健作: "Python广告数据挖掘与分析实战", 31 March 2021, 机械工业出版社 , pages: 190 - 191 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112699302A (en) * | 2021-01-05 | 2021-04-23 | 汕头大学 | Real-time recommendation method based on geodesic distance |
| CN112699302B (en) * | 2021-01-05 | 2023-07-04 | 汕头大学 | Real-time recommendation method based on geodesic distance |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108288229B (en) | User portrait construction method | |
| US9152969B2 (en) | Recommendation ranking system with distrust | |
| US8214375B2 (en) | Manual and automatic techniques for finding similar users | |
| US20100250526A1 (en) | Search System that Uses Semantic Constructs Defined by Your Social Network | |
| CN110020188A (en) | Recommended based on the Global Vector of implicit interactions and profile data | |
| Skopik et al. | Start trusting strangers? bootstrapping and prediction of trust | |
| WO2016196526A1 (en) | Viewport-based implicit feedback | |
| CN112860989B (en) | Course recommendation method and device, computer equipment and storage medium | |
| US20150073932A1 (en) | Strength Based Modeling For Recommendation System | |
| WO2020003109A1 (en) | Facet-based query refinement based on multiple query interpretations | |
| CN109542757A (en) | Interface testing environment determines method, apparatus, electronic equipment and storage medium | |
| CN114862140A (en) | Behavior analysis-based potential evaluation method, device, equipment and storage medium | |
| CN113704608A (en) | Personalized item recommendation method and device, electronic equipment and storage medium | |
| Brambilla et al. | An explorative approach for crowdsourcing tasks design | |
| CN109800147A (en) | A kind of test cases generation method and terminal device | |
| US20140244741A1 (en) | Computer-Implemented System And Method For Context-Based APP Searching And APP Use Insights | |
| US20130159293A1 (en) | Generating a supplemental description of an entity | |
| CN112328752B (en) | Course recommendation method and device based on search content, computer equipment and medium | |
| CN111737258B (en) | Method and device for representing, recording and automatically generating product design scheme | |
| CN114357293A (en) | Object recommendation method and device, electronic equipment and computer-readable storage medium | |
| CN107295074A (en) | It is a kind of to realize the method and apparatus that cloud resource is shown | |
| CN114676272A (en) | Information processing method, device and equipment of multimedia resource and storage medium | |
| CN111046300A (en) | Method and device for determining crowd attributes of users | |
| CN114398501B (en) | Multimedia resource grouping method, device, equipment and storage medium | |
| CN110264333A (en) | A kind of risk rule determines method and apparatus |
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 |