CN115713386A - Multi-source information fusion commodity recommendation method and system - Google Patents
Multi-source information fusion commodity recommendation method and system Download PDFInfo
- Publication number
- CN115713386A CN115713386A CN202211483066.1A CN202211483066A CN115713386A CN 115713386 A CN115713386 A CN 115713386A CN 202211483066 A CN202211483066 A CN 202211483066A CN 115713386 A CN115713386 A CN 115713386A
- Authority
- CN
- China
- Prior art keywords
- user
- project
- item
- potential
- information
- 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
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention belongs to the field of recommendation systems, and provides a multi-source information fusion commodity recommendation method and system, which comprises the steps of obtaining a user-item diagram and user comment information to perform user modeling, and determining a user potential factor; acquiring a user-project diagram and project comment information to perform project modeling, and determining potential factors of a project; splicing the user potential factor and the project potential factor to obtain a user-project potential factor; and carrying out scoring prediction based on the user-item potential factors to obtain a commodity scoring prediction result, and recommending commodities to the user by using the commodity scoring prediction result. The invention uses GNN and CNN to process interaction information and comment information respectively and then carries out feature fusion and prediction scoring.
Description
Technical Field
The invention belongs to the technical field of recommendation systems, and particularly relates to a multi-source information fusion commodity recommendation method and system.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
In recent years, recommendation systems are more and more widely applied and favored by more users, and a lot of researchers are attracted to the application, so that academic papers on recommendation systems and various books on recommendation system technologies are rapidly growing, various recommendation models are endlessly developed, and most model principles are based on a basic method of collaborative filtering, which makes the model based on collaborative filtering a very successful technical solution so far. The collaborative filtering algorithm carries out recommendation by utilizing the interactive information between the user and the project, and the method is simple and effective, but faces the problems of data sparsity and cold start. Fusing the auxiliary information of users or projects for the two problems is an effective solution. The auxiliary information comprises comment text, social networks, attributes and the like, wherein the comment information further comprises semantic information rich in users and items. Therefore, how to effectively utilize the comment auxiliary information and improve the performance of the recommendation system is an important problem.
In recent years, most comment recommendation models only use comment information of users and items as recommendations, which cannot sufficiently extract potential information of the users and the items.
Disclosure of Invention
In order to solve the problems, the invention provides a multi-source information fusion commodity recommendation method and system. The user project graph is aggregated through the graph network, and the potential characteristic information of the user and the article can be more fully expressed as the supplement of the comment information.
According to some embodiments, a first aspect of the present invention provides a multi-source information fused commodity recommendation method, which adopts the following technical solutions:
a multi-source information fused commodity recommendation method comprises the following steps:
acquiring a user-project graph and user comment information to perform user modeling, and determining a user potential factor;
acquiring a user-project diagram and project comment information to perform project modeling, and determining potential factors of a project;
splicing the user potential factor and the project potential factor to obtain a user-project potential factor;
and performing rating prediction based on the user-item latent factor to obtain a commodity rating prediction result, and recommending commodities to the user by using the commodity rating prediction result.
Further, the obtaining of the user-project graph and the user comment information for user modeling and determining the user potential factor includes:
text information coding is carried out based on the user comment information, and user comment characteristics are obtained;
acquiring a user-project graph to aggregate projects and learn user potential factors of a project space;
and carrying out feature concatenation on the user comment features and the potential factors of the project space to determine the potential factors of the user.
Further, the encoding of the text information based on the user comment information to obtain the user comment features includes:
converting the user comment information into an embedding matrix by using a word embedding function;
performing convolution operation on the embedded matrix based on the convolution layer to generate embedded matrix characteristics;
and the embedded matrix features pass through a maximum pooling layer and a fusion connection layer to obtain user comment features.
Further, the acquiring the user-project graph to aggregate the projects and learn the user latent factors of the project space includes:
determining attention weights of the user and the item by using an attention network based on the interactive item embedding and the user embedding in the user-item diagram;
user latent factors of the project space are learned by considering the projects with which the user interface has interacted in conjunction with the attention weights of the user and the projects.
Further, the obtaining of the user-project graph and the project comment information for project modeling and determining the potential factors of the project comprise:
coding text information based on the item comment information to obtain item comment characteristics;
aggregating information from a set of users that interacted with each item;
determining importance weights for distinguishing users by using an attention network according to interactive user embedding and project embedding;
and combining the important weight of the user, and performing feature series connection on the aggregated information and the comment features of the project to obtain potential factors of the project.
Further, the user potential factor and the project potential factor are spliced to obtain a user-project potential factor, which specifically includes:
wherein the item latent factor z j User latent factor h i 。
Further, the scoring prediction is performed based on the user-item potential factor, a commodity scoring prediction result is obtained by the user-item potential factor through MLP, and commodity recommendation is performed on the user by using the commodity scoring prediction result, specifically:
g 2 ＝(W 2 * 1 + 2 )
where l is the hidden layer index, r i ′ j Is user u i For u is paired j The prediction score of (a);
recommending the item with the prediction score at the top to the user.
According to some embodiments, a second aspect of the present invention provides a multi-source information fusion commodity recommendation system, which adopts the following technical solutions:
a multi-source information fused commodity recommendation system comprises:
the user potential factor determining module is configured to acquire a user-project graph and user comment information for user modeling and determine a potential factor of a user;
the project potential factor determining module is configured to acquire a user-project diagram and project comment information to perform project modeling and determine a potential factor of a project;
the data series connection module is configured to splice the potential factors of the user and the potential factors of the project to obtain user-project potential factors;
and the commodity recommending module is configured to perform grading prediction based on the user-item potential factor to obtain a commodity grading prediction result, and perform commodity recommendation on the user by using the commodity grading prediction result.
According to some embodiments, a third aspect of the invention provides a computer-readable storage medium.
A computer-readable storage medium, on which a computer program is stored, which when executed by a processor implements the steps in a multi-source information-fused merchandise recommendation method according to the first aspect.
According to some embodiments, a fourth aspect of the invention provides a computer apparatus.
A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor executes the program to implement the steps of the multi-source information fusion merchandise recommendation method according to the first aspect.
Compared with the prior art, the invention has the beneficial effects that:
1) Compared with the current recommendation model based on the comment information, the recommendation model is supplemented by the interactive data between the articles and the users, so that the comment prediction error is reduced.
(2) The method creatively combines the interaction information of the user project processed by the GNN and the CNN processing comment information for recommendation, and provides an idea for multi-source data fusion recommendation later.
(3) The model framework has expandability and high flexibility, and a fusion model suitable for the model framework can be designed according to a specific service scene.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, are included to provide a further understanding of the invention, and are incorporated in and constitute a part of this specification, illustrate exemplary embodiments of the invention and together with the description serve to explain the invention and not to limit the invention.
FIG. 1 is a diagram of a user modeling network architecture according to an embodiment of the present invention;
FIG. 2 is a user-item interaction diagram according to an embodiment of the invention;
FIG. 3 is a diagram of a project modeling network architecture according to an embodiment of the present invention;
fig. 4 is a diagram of a structure of a model for comment prediction according to an embodiment of the present invention.
Detailed Description
The invention is further described with reference to the following figures and examples.
It is to be understood that the following detailed description is exemplary and is intended to provide further explanation of the invention as claimed. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments according to the invention. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
The embodiments and features of the embodiments of the invention may be combined with each other without conflict.
Example one
As shown in fig. 1, the embodiment provides a multi-source information fused commodity recommendation method, and the embodiment is exemplified by applying the method to a server, and it can be understood that the method can also be applied to a terminal, and can also be applied to a system including the terminal and a server, and is implemented by interaction between the terminal and the server. The server may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server providing basic cloud computing services such as a cloud service, a cloud database, cloud computing, a cloud function, cloud storage, a network server, cloud communication, middleware service, a domain name service, a security service CDN, a big data and artificial intelligence platform, and the like. The terminal may be, but is not limited to, a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, and the like. The terminal and the server may be directly or indirectly connected through wired or wireless communication, and the application is not limited herein. In this embodiment, the method includes the steps of:
a multi-source information fused commodity recommendation method comprises the following steps:
acquiring a user-project graph and user comment information to perform user modeling, and determining a user potential factor;
acquiring a user-project graph and project comment information to perform project modeling, and determining potential factors of the project;
splicing the user potential factor and the project potential factor to obtain a user-project potential factor;
and carrying out scoring prediction based on the user-item potential factors to obtain a commodity scoring prediction result, and recommending commodities to the user by using the commodity scoring prediction result.
Specifically, the process of the method described in this embodiment specifically includes:
the model is divided into 3 modules, namely user modeling, project modeling and grading modeling.
User modeling
The first part is user modeling, i.e. learning the underlying factors of the user. It can understand the user through the interaction between the user and the project in the user-project diagram and extract the comment information of the user, as shown in fig. 1.
Comment information coding
The method adopts a coding mode of a DeepConn model for comment text information, firstly defines a word embedding function f, namely M → R d Where M represents a dictionary, each word in the comment may be mapped to a d-dimensional dense vector. Each comment is a word sequence of length L. Each word is then mapped to a d-dimensional dense vector. Each comment is a word sequence of length L, and then each comment is converted into an embedding matrix by a word embedding function. The next layer in the embedding layer is a convolutional layer, which contains n neurons, each associated with a filter for convolution operations to produce new features.
Wherein, the first and the second end of the pipe are connected with each other,
is an embedding matrix of the i-th comment information of the user u, K j Is the jth convolution kernel, b j Is an offset term, z j Is the feature produced by the jth neuron. The most important features in each feature map are then captured through a max pooling operation. The fusion connection operation connects the features generated by n neurons together to obtain feature information of a comment, as shown in the formula:
and obtaining the comment representation of the user u by aggregating the representation of each comment of the user u, wherein the formula is as follows:
and (4) the result of the maximum pooling layer passes through a full-connection layer with a weight matrix W and an offset g to obtain the final comment feature representation.
x u ＝ ′ (*O u +)
Item aggregation
The user project diagram is a diagram representing interaction information between a user and a project, as shown in fig. 2. Let G (B, E) represent a user-item bipartite graph, with B = B U ∩B V By the user's vertex set B U And item vertex set B. For each doublet (u, v) in the dataset, there is a corresponding edge e (b) u ,b v ) Wherein b is u ∈B U Is the vertex corresponding to user u, b v ∈B V Is the vertex corresponding to item v.
The purpose of item aggregation is to learn the user latent factors of the item space by considering the items with which the user interface has interacted. Expressed mathematically:
wherein C (i) is user u i Set of interacted items, q a For interactive item embedding, aggre items The aggregation function is the mean operator for the item. δ is the nonlinear activation function, and W and b are the weights and bias ranks of the neural network.
In the formula, a i Is fixed to
Then, a mean value a of two-layer attention network parameterized accumulation is adopted i . The input to the attention network is the interactive item embedding q a And user embedding e u . In form, the attention network is defined as:
the attention scores are normalized by using a Softmax function to obtain final attention weights, which can be interpreted as the contribution of interaction to the project-space user potential factors of the user:
feature fusion
The first two parts obtain information from different angles of the user from the user-project diagram and comment information. Then 2 characteristics are input into the MLP network in series to obtain the final user potential factor h i . User latency factor h i Is defined as follows:
wherein l is a hidden layer index.
Project modeling
The structure of the item modeling of the second part is consistent with the user modeling, as shown in FIG. 3. Comment information modeling in item modeling is relative to comment information modeling part in user modeling, and comment information of a user is replaced by comment information of an item, so that comment characteristic representation x of the item is obtained v . And changing item aggregation in user modeling to user replacement. The process of user aggregation is as follows:
the purpose of user aggregation is to learn the item latent factors of the user space by considering the users with whom the item interface has interacted. Likewise, v for each item j We need to be derived from v j The information is aggregated in the set of interacted users (denoted as B (j)). Is shown mathematically
Then, the attention system of the two-layer neural network is utilized to obtain the important point for distinguishing the usersSex weights
Embedding p for interactive users with input at the attention level t Embedding with items e i Is made by splicing
The formula is as follows:
the first two parts acquire information of projects from different angles from a user-project diagram and comment information, and then 2 features are serially connected and input into an MLP network to obtain final project potential factors. Item latent factor z j Is defined as:
wherein l is a hidden layer index.
Scoring prediction
With the potential factors of the user and the item being (h) i And z j ) Then connecting them
Score prediction was performed as follows:
wherein, l is a hidden layer cableFrom r' ij Is user u i For u is paired j The prediction score of (1). The model structure of the comment prediction is shown in fig. 4.
Model training
Because the method predicts the scores, the mean square error common in score prediction is selected as a loss function, as follows:
where | | | is the number of scores observed, and r ij Is truly scored on item j by user i. In order to optimize the objective function, RMSprop was used as the optimizer. One training instance is randomly selected at a time and each model parameter is updated in the direction of its negative gradient.
Example two
The embodiment provides a commodity recommendation system of multisource information fusion, includes:
the user potential factor determining module is configured to acquire a user-project graph and user comment information for user modeling and determine a potential factor of a user;
the project potential factor determining module is configured to acquire a user-project diagram and project comment information to perform project modeling and determine a potential factor of a project;
the data series connection module is configured to splice the potential factors of the user and the potential factors of the project to obtain user-project potential factors;
and the commodity recommending module is configured to perform grading prediction based on the user-item potential factor to obtain a commodity grading prediction result, and perform commodity recommendation on the user by using the commodity grading prediction result.
The modules are the same as the corresponding steps in the implementation example and application scenarios, but are not limited to the disclosure of the first embodiment. It should be noted that the modules described above as part of a system may be implemented in a computer system such as a set of computer-executable instructions.
In the foregoing embodiments, the descriptions of the embodiments have different emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments.
The proposed system can be implemented in other ways. For example, the above-described system embodiments are merely illustrative, and for example, the division of the above-described modules is merely a logical functional division, and in actual implementation, there may be another division, for example, a plurality of modules may be combined or may be integrated into another system, or some features may be omitted, or not executed.
EXAMPLE III
The present embodiment provides a computer-readable storage medium, on which a computer program is stored, which when executed by a processor implements the steps in a multi-source information-fused product recommendation method as described in the first embodiment above.
Example four
The embodiment provides a computer device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor executes the program to implement the steps in the multi-source information fusion merchandise recommendation method according to the first embodiment.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of a hardware embodiment, a software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, which can be stored in a computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. The storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), or the like.
Although the embodiments of the present invention have been described with reference to the accompanying drawings, it is not intended to limit the scope of the present invention, and it should be understood by those skilled in the art that various modifications and variations can be made without inventive efforts by those skilled in the art based on the technical solution of the present invention.
Claims (10)
1. A multi-source information fused commodity recommendation method is characterized by comprising the following steps:
acquiring a user-project graph and user comment information for user modeling, and determining a user potential factor;
acquiring a user-project diagram and project comment information to perform project modeling, and determining potential factors of a project;
splicing the user potential factor and the project potential factor to obtain a user-project potential factor;
and carrying out scoring prediction based on the user-item potential factors to obtain a commodity scoring prediction result, and recommending commodities to the user by using the commodity scoring prediction result.
2. The multi-source information fused commodity recommendation method according to claim 1, wherein the obtaining of the user-item map and the user comment information for user modeling and determining the user latent factor comprises:
text information coding is carried out based on the user comment information, and user comment characteristics are obtained;
acquiring a user-project graph to aggregate projects and learn user potential factors of a project space;
and carrying out feature concatenation on the user comment features and the potential factors of the project space to determine the potential factors of the user.
3. The multi-source information fused commodity recommendation method according to claim 2, wherein the encoding of text information based on user comment information to obtain user comment features comprises:
converting the user comment information into an embedded matrix by using a word embedding function;
performing convolution operation on the embedded matrix based on the convolution layer to generate embedded matrix characteristics;
and the embedded matrix features pass through a maximum pooling layer and a fusion connection layer to obtain user comment features.
4. The multi-source information fused commodity recommendation method according to claim 2, wherein the acquiring of the user-item graph to aggregate items and learning of the user latent factors of the item space comprises:
determining attention weights of the user and the item by using an attention network based on the interactive item embedding and the user embedding in the user-item diagram;
user latent factors of the project space are learned by considering the projects with which the user interface has interacted in conjunction with the attention weights of the user and the projects.
5. The multi-source information fused commodity recommendation method according to claim 1, wherein the obtaining of the user-item diagram and the item review information for item modeling and determining the potential factors of the item comprises:
coding text information based on the item comment information to obtain item comment characteristics;
aggregating information from a set of users who interacted with each item;
determining importance weights for distinguishing users by using an attention network according to interactive user embedding and project embedding;
and combining the important weight of the user, and performing feature series connection on the aggregated information and the comment features of the project to obtain potential factors of the project.
6. The multi-source information fused commodity recommendation method according to claim 1, wherein the user potential factor and the item potential factor are spliced to obtain a user-item potential factor, and the method specifically comprises the following steps:
wherein the item latent factor z j User latent factor h i 。
7. The multi-source information fusion commodity recommendation method according to claim 1, wherein the scoring prediction is performed based on the user-item latent factor, a commodity scoring prediction result is obtained from the user-item latent factor through MLP, and commodity recommendation is performed on the user by using the commodity scoring prediction result, specifically:
r′ ij ＝W T *g l-1
wherein l is a hidden layer index, r' ij Is user u i For u is paired j The prediction score of (a);
recommending the item with the prediction score at the top to the user.
8. A multi-source information fused commodity recommendation system is characterized by comprising:
the user potential factor determining module is configured to acquire a user-project graph and user comment information for user modeling and determine a potential factor of a user;
the project potential factor determining module is configured to acquire a user-project diagram and project comment information to perform project modeling and determine a potential factor of a project;
the data series connection module is configured to splice the potential factors of the user and the potential factors of the project to obtain user-project potential factors;
and the commodity recommending module is configured to perform grading prediction based on the user-item potential factor to obtain a commodity grading prediction result, and perform commodity recommendation on the user by using the commodity grading prediction result.
9. A computer-readable storage medium, on which a computer program is stored, wherein the program, when executed by a processor, implements the steps in a multi-source information-fused product recommendation method according to any one of claims 1-7.
10. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the multi-source information fusion merchandise recommendation method according to any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211483066.1A CN115713386A (en) | 2022-11-24 | 2022-11-24 | Multi-source information fusion commodity recommendation method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211483066.1A CN115713386A (en) | 2022-11-24 | 2022-11-24 | Multi-source information fusion commodity recommendation method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115713386A true CN115713386A (en) | 2023-02-24 |
Family
ID=85234876
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211483066.1A Pending CN115713386A (en) | 2022-11-24 | 2022-11-24 | Multi-source information fusion commodity recommendation method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115713386A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117994007A (en) * | 2024-04-03 | 2024-05-07 | 山东科技大学 | Social recommendation method based on multi-view fusion heterogeneous graph neural network |
-
2022
- 2022-11-24 CN CN202211483066.1A patent/CN115713386A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117994007A (en) * | 2024-04-03 | 2024-05-07 | 山东科技大学 | Social recommendation method based on multi-view fusion heterogeneous graph neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109739939A (en) | The data fusion method and device of knowledge mapping | |
| US11227190B1 (en) | Graph neural network training methods and systems | |
| Zliobaite et al. | Next challenges for adaptive learning systems | |
| WO2022142006A1 (en) | Semantic recognition-based verbal skill recommendation method and apparatus, device, and storage medium | |
| CN110929164A (en) | Interest point recommendation method based on user dynamic preference and attention mechanism | |
| CN111444340A (en) | Text classification and recommendation method, device, equipment and storage medium | |
| WO2023065859A1 (en) | Item recommendation method and apparatus, and storage medium | |
| US20220253722A1 (en) | Recommendation system with adaptive thresholds for neighborhood selection | |
| CN107741976A (en) | Intelligent response method, apparatus, medium and electronic equipment | |
| US11423307B2 (en) | Taxonomy construction via graph-based cross-domain knowledge transfer | |
| CN111353033A (en) | Method and system for training text similarity model | |
| CN111324773A (en) | Background music construction method and device, electronic equipment and storage medium | |
| Goyal et al. | Introduction to natural language processing and deep learning | |
| CN115713386A (en) | Multi-source information fusion commodity recommendation method and system | |
| CN114117048A (en) | Text classification method and device, computer equipment and storage medium | |
| CN113255327B (en) | Text processing method and device, electronic equipment and computer readable storage medium | |
| CN114330476A (en) | Model training method for media content recognition and media content recognition method | |
| CN114330514A (en) | Data reconstruction method and system based on depth features and gradient information | |
| CN113837307A (en) | Data similarity calculation method and device, readable medium and electronic equipment | |
| CN117291722A (en) | Object management method, related device and computer readable medium | |
| CN111708880A (en) | System and method for identifying class cluster | |
| CN116756281A (en) | Knowledge question-answering method, device, equipment and medium | |
| CN116957128A (en) | Service index prediction method, device, equipment and storage medium | |
| CN115344794A (en) | Scenic spot recommendation method based on knowledge map semantic embedding | |
| CN115510193A (en) | Query result vectorization method, query result determination method and related device |
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 |