CN112395506A - Information recommendation method and device, electronic equipment and storage medium - Google Patents
Information recommendation method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN112395506A CN112395506A CN202011408670.9A CN202011408670A CN112395506A CN 112395506 A CN112395506 A CN 112395506A CN 202011408670 A CN202011408670 A CN 202011408670A CN 112395506 A CN112395506 A CN 112395506A
- Authority
- CN
- China
- Prior art keywords
- information
- user behavior
- vector
- recommendation
- user
- 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
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
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- 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
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Business, Economics & Management (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Software Systems (AREA)
- Bioinformatics & Computational Biology (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Animal Behavior & Ethology (AREA)
- Biophysics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Finance (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- Technology Law (AREA)
- General Business, Economics & Management (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses an information recommendation method, an information recommendation device, electronic equipment and a storage medium, wherein the method comprises the following steps: constructing a financial knowledge map according to entities of a preset financial data set; determining user behavior characteristics based on the financial knowledge graph and user information browsing records; and selecting target information for recommendation according to the user behavior characteristics. The embodiment of the invention identifies the context knowledge of the information through the financial knowledge map, enriches the semantic representation of the information, improves the accuracy of the personalized information recommendation of the user and can enhance the quality of the information recommendation.
Description
Technical Field
The embodiment of the invention relates to the technical field of computer application, in particular to an information recommendation method, an information recommendation device, electronic equipment and a storage medium.
Background
With the development of the internet and the popularization of mobile terminals, we are experiencing an information explosion era. How to find out the interesting content from the massive information is not only the expectation of each user, but also the competitive point of interest among internet products. In the internet era with the flow being at the top, each internet product can occupy more user markets and has more flow only by accurately positioning the preference of the user, customizing personalized browsing content for the user and improving the use experience of the user. Therefore, how to perform accurate user representation and recommend desired information to the user becomes a problem to be solved urgently in the field.
In the prior art, a unit for recommending contents is easily caused by simply recommending a user through historical reading contents of the user and information of the user, only information related to the historical reading contents is recommended for the user, for example, the user browses information about science and news in the past, the user can take the information and also be interested in a new four-dimensional map which is a technological innovation block, but the existing method cannot recommend the new information of the four-dimensional map for the user.
Disclosure of Invention
The invention provides an information recommendation method, an information recommendation device, electronic equipment and a storage medium, which increase external knowledge through a knowledge map, improve the richness of information recommendation, enhance the quality of information recommendation and improve the use experience degree of a user.
In a first aspect, an embodiment of the present invention provides an information recommendation method, where the method includes:
constructing a financial knowledge map according to entities of a preset financial data set;
determining user behavior characteristics based on the financial knowledge graph and user information browsing records;
and selecting target information for recommendation according to the user behavior characteristics.
In a second aspect, an embodiment of the present invention further provides an information recommendation apparatus, where the apparatus includes:
the map construction model is used for constructing a financial knowledge map according to the entities of the preset financial data set;
the user characteristic module is used for determining user behavior characteristics based on the financial knowledge map and the user information browsing record;
and the information recommendation module is used for selecting target information to recommend according to the user behavior characteristics.
In a third aspect, an embodiment of the present invention further provides an electronic device, where the electronic device includes:
one or more processors;
a memory for storing one or more programs,
when the one or more programs are executed by the one or more processors, the one or more processors implement the information recommendation method according to any one of the embodiments of the present invention.
In a fourth aspect, the present invention further provides a computer-readable storage medium, on which a computer program is stored, where the computer program is executed by a processor to implement the method according to any one of the embodiments of the present invention.
According to the embodiment of the invention, the financial knowledge map is established through the entity of the preset financial data set, the user behavior characteristics are determined through the financial knowledge map and the user information browsing record, and the target information is selected for recommendation based on the user behavior characteristics, so that the individuation of information recommendation is improved, the information recommendation range is enriched, and the information recommendation quality can be enhanced.
Drawings
FIG. 1 is a flowchart illustrating an information recommendation method according to an embodiment of the present invention;
FIG. 2 is a flowchart illustrating an information recommendation method according to a second embodiment of the present invention;
FIG. 3 is an exemplary diagram of a financial knowledge-graph as provided in a second embodiment of the invention;
fig. 4 is a schematic structural diagram of a feature extraction model provided in the second embodiment of the present invention;
FIG. 5 is a schematic structural diagram of an information recommendation apparatus according to a third embodiment of the present invention;
fig. 6 is a schematic structural diagram of an electronic device according to a fourth embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be noted that, for convenience of description, only a part of the structures related to the present invention, not all of the structures, are shown in the drawings, and furthermore, embodiments of the present invention and features of the embodiments may be combined with each other without conflict.
Example one
Fig. 1 is a flowchart of an information recommendation method in an embodiment of the present invention, where the embodiment is applicable to personalized recommendation of information, the method may be executed by an information recommendation apparatus, and the apparatus may be implemented in a hardware and/or software manner, and referring to fig. 1, the method provided in the embodiment of the present invention may generally include the following steps:
The preset financial data set may be a data set collected in advance, and may include company information, industry information, and the like, and the preset financial data set may specifically obtain industry classification information of all listed companies in Shanghai and deep market from Wand (wind). The financial knowledge graph may be information for mining and displaying financial domain knowledge, and each edge of the financial knowledge graph exists in the form of an entity-relationship triple (h, r, t), where h, r, t respectively represent a financial entity, a relationship, a tail entity, and the like.
In the embodiment of the invention, text recognition can be performed on the preset financial data set, and entities in the preset financial data set, such as company names and industry names, can be obtained. The knowledge graph can be constructed according to the entities and the incidence relation among the entities. For example, each entity may be used as a vertex in the knowledge graph, and the association relationship between the corresponding entities may be converted into a connection line between the vertices.
And step 120, determining user behavior characteristics based on the financial knowledge graph and the user information browsing record.
The user information browsing record may be a record generated by the user browsing information in the past, and the user information browsing record may include a title, a content, and a browsing time browsed by the user. The user behavior feature may be a habit feature reflecting the user browsing the financial information, and the user behavior feature may be specifically a vector, and the vector may include the financial information for browsing like or browsing dislike.
Specifically, the user information browsing records may be subjected to feature extraction based on the financial knowledge graph, the extracted features may be used as the user behavior features, for example, the record with the largest browsing frequency or the longest browsing time is extracted from the user information browsing records, the corresponding entity may be searched in the financial knowledge graph according to the record, and the financial information corresponding to the entity may be used as the user behavior features. The financial knowledge map and the user browsing records can be input into the neural network model together for feature extraction, and the extracted result can be used as the user behavior feature, wherein the neural network model can be generated through training of the financial knowledge map and the information browsing records of massive users in advance.
And step 130, selecting target information to recommend according to the user behavior characteristics.
The target information can be information corresponding to the user behavior characteristics, the target information can be generated by selecting from an information set to be recommended, and the selection principle can be the user behavior characteristics.
Specifically, one or more pieces of information in the information set to be recommended can be selected as target information for recommendation based on the information content included in the user behavior characteristics. The recommendation probabilities corresponding to the information in the information set to be recommended can be predicted according to the user behavior characteristics, and one or more information can be selected as target information to be recommended according to the sequence of the recommendation probabilities from large to small.
According to the embodiment of the invention, the entity of the preset financial data set is extracted to establish the financial knowledge map, the user information browsing record is analyzed based on the financial knowledge map to obtain the user behavior characteristics, and the target information is selected for recommendation based on the user behavior characteristics, so that the individuation of information recommendation is improved, the information recommendation range is enriched, and the information recommendation quality can be enhanced.
Example two
Fig. 2 is a flowchart of an information recommendation method in a second embodiment of the present invention, which is embodied on the basis of the above-mentioned embodiment of the present invention, and referring to fig. 2, the method provided by the embodiment of the present invention specifically includes the following steps:
The noun text can be a vocabulary with actual meanings, and can include place name, person name, organization name, company name, industry name and the like, the entity can be a component of a financial knowledge graph, the entity and the noun text can have a corresponding relationship, one or more noun texts can correspond to one entity, for example, the science news and technology news and the four-dimensional map can correspond to the same entity, and the entity can represent a technological innovation board in the financial knowledge graph.
In the embodiment of the invention, the word segmentation operation can be carried out on the text in the preset financial data set to obtain different word segments, and the word segments can contain text vocabularies without actual meanings, so that the obtained word segments can be directly identified by utilizing a word bank to obtain noun texts with actual meanings as entities. Furthermore, due to the diversity and ambiguity of natural semantic expressions, noun texts recognized from the preset financial data set may have a plurality of noun texts with different expressions for the same entity, the similarity of each noun text may be calculated through a semantic fuzzy machine learning model, and the noun texts with the same semantics may be mapped to the same entity.
And step 220, reducing the dimensionality of the entity according to the preset graph representation learning processing operation to obtain a financial knowledge graph.
The preset graph represents the operation that the learning operation can embed the view into a low-dimensional and dense vector space, and the processed entity can store the internal structure information of the reason.
In the embodiment of the invention, the dimensionality of the entity can be reduced according to the preset graph representation learning processing operation to be embedded into the low-dimensional vector space, the low-dimensional vector space embedded with the entity can be used as the financial knowledge graph, so that a continuous space vector is formed after dimensionality reduction, the data processing amount in the subsequent information recommendation process is reduced, and each edge in the financial knowledge graph can be represented by an entity-relation triple (h, r, t). The preset graph represents learning processing operation, wherein the learning processing operation can comprise two types of processing operation based on Trans and path. Furthermore, the financial knowledge graph can comprise two types of entities of companies and industries and a type of relation of the companies belonging to the industries, the method based on the Trans is more suitable for graph representation with more types of relations, and the method based on the path is more suitable for graph representation with similar attributes for entities with closer distances. For example, the dimension of the entity vector is set to 50, and the financial knowledge graph formed by performing dimension reduction on the obtained entity vector can be shown in fig. 3, wherein entities in different industries are represented by icons in different shapes, so that it can be seen that entities in the same industry are all gathered in the same cluster, and entities in different industries can also be clearly distinguished. For example, the cluster that is clustered in the lower left corner of the figure is a company in the mechanical equipment industry, and the cluster that is clustered in the middle is a company in the pharmaceutical and biological industry.
And step 230, acquiring a map vector corresponding to the financial knowledge map and a user behavior vector corresponding to the user information browsing record.
In particular, the financial knowledge-graph may be represented in the form of a vector, for example, the entity distribution in the financial knowledge-graph may be used as an element of the vector. Information titles or information labels in the user information browsing records can be extracted, and the acquired titles or labels can be used as constituent elements in the user behavior vectors.
And 240, inputting the graph spectrum vector and the user behavior vector into a feature extraction model to obtain the user behavior feature, wherein the feature extraction model is generated by training of a historical browsing data set.
The feature extraction model may be a model for extracting the reading habit features of the user, specifically, a convolutional neural network model, and the feature extraction model may include one or more input channels. The user behavior characteristics may be specifically characteristics of information read by the user, such as a favorite information type or a least favorite information type, the historical browsing data set may be a data set including information browsing records of one or more users, a data scale of the historical browsing data set may satisfy a threshold condition, and the larger the data scale of the historical browsing data set is, the higher accuracy of extracting the user behavior characteristics by the feature extraction model generated through training of the historical browsing data may be. In the embodiment of the invention, the historical browsing data set can be used as a positive sample of the training set, a data set is randomly generated as a negative sample, and the feature extraction model can be trained through the positive sample and the negative sample until the output result of the extracted model meets the requirement.
In the embodiment of the invention, the acquired map vector and the user behavior vector can be input into the feature extraction model, the map vector and the user behavior vector can be fused and input into the feature extraction model as one input vector, the map vector and the user behavior vector can also be respectively input into the feature extraction model as one vector, and the user features can be acquired by processing the map vector and the user behavior vector through the feature extraction model.
And step 250, determining recommendation probability corresponding to the information to be recommended based on the user behavior characteristics.
The information to be recommended may be a preset information data set, and the information to be recommended may include one or more pieces of information.
Specifically, the recommendation probability of each piece of recommendation information in the information to be recommended can be determined according to the user behavior characteristics, the larger the recommendation probability is, the more likely the corresponding recommendation information is to be read by the user, illustratively, a long-short term memory network model can be trained according to the information reading history of a large number of users and the information to be recommended, the long-short term memory network model is used for determining the recommendation probability corresponding to the information to be recommended, the user behavior characteristics and the information to be recommended can be input into the long-short term memory network model to obtain the corresponding recommendation probability, it can be understood that the information to be recommended can include a plurality of information, and the user behavior characteristics and the information to be recommended can be input into the long-short term memory network model in a vector manner to obtain the recommendation probability corresponding to each piece of recommendation information to be recommended.
And step 260, selecting the corresponding information to be recommended as target information to recommend according to the recommendation probability.
In the embodiment of the invention, the corresponding information to be recommended can be selected as the target information according to the recommendation probability and recommended to the user, and the information to be recommended with larger recommendation probability can be selected as the target information. In another embodiment, one or more pieces of information to be recommended with a recommendation probability greater than a threshold value may be selected as target information to be recommended to the user together.
According to the embodiment of the invention, noun texts in a preset financial data set are identified as entities, the dimensionality of the entities is reduced and a financial knowledge map is constructed according to a preset map representation, map vectors corresponding to the financial knowledge map and user behavior vectors corresponding to user information browsing records are obtained and input into a feature extraction model to obtain user behavior features, the recommendation probability of information to be recommended is determined based on the user behavior features, target information is obtained according to the recommendation probability for recommendation, the personalization degree of user information recommendation is improved, the information recommendation range can be enriched, and the experience degree of information reading of users is enhanced.
Further, on the basis of the embodiment of the present invention, the preset map indicates that the learning process at least includes one of the following steps: graph representation learning based on TransE, graph representation learning based on TransD, graph representation learning based on TransR, Deepwalk graph representation learning and Node2vec graph representation learning.
In the embodiment of the present invention, the graph representation based on the TransE learns the entity-relationship triplet (h, r, t) corresponding to one edge of the financial knowledge graph, where h, r, t respectively represent the head entity, the relationship and the tail entity, and it is assumed that (h, r, t) satisfies h + r ≈ t, where r, t respectively represent the head entity, the relationship and the vector representation of the tail entity, i.e., the triplet satisfying the (h, r, t) relationship corresponds to the smaller score function of the following formula, and the larger score function otherwise.
And the graph representation learning based on the TransD assumes that a mapped hyperplane exists for each relationship of different types of entities, and each entity is firstly mapped to the hyperplane of the corresponding relationship by the following formula:
wherein, wrIs the normal vector of the r relation hyperplane. The triples after mapping satisfy the score function of TransE, namely:
graph representation learning based on TransR assumes that a relationship space exists for each of different types of entity relationships, and the entities are mapped through a mapping matrix MrIs mapped to the corresponding space:
hr＝hMr,tr＝tMr
the mapped vector also satisfies the score function of TransE, namely:
further, Deepwalk represents that learning can generate a sequence of nodes by using the simplest random walk mode, namely the mode of selecting the next node is uniformly and randomly distributed, the nodes in the sequence can be regarded as words in a text, and a Skip-Gram model is used for training vectors of the nodes.
The Node2vec graph shows that learning can be based on changing the generation mode of the Deepwalk random walk sequence, and width-first search and depth-first search are introduced into the generation process of the random walk sequence by introducing two parameters p and q.
Further, on the basis of the above embodiment of the present invention, the identifying noun text in the preset financial data set as an entity includes: determining the similarity of each noun text through a preset text similarity model; and mapping noun texts with the same semantics to the same entity according to the similarity.
In the embodiment of the present invention, the preset text similarity model may be a deep learning model that determines the similarity between noun texts, and the model may be generated through massive text vocabulary training. Noun texts identified in a preset financial data set can be input into a preset text similarity model, an output result is obtained, the output result can be a result vector, each element can correspond to one noun text, and the noun texts with the same value or the difference value smaller than a fixed threshold value in the result vector can be mapped to the same entity as the same type of text.
Further, on the basis of the above embodiment of the present invention, the inputting the atlas vector and the user behavior vector into a feature extraction model to obtain a user behavior feature includes:
inputting the map vector and the user behavior vector into a feature extraction model respectively, wherein the map vector and the user behavior vector correspond to one input channel of the feature extraction model respectively; and performing convolution and pooling operation on the atlas vector and the user behavior vector in the feature extraction model to obtain an information vector as the user behavior feature.
In the embodiment of the invention, the feature extraction model can be provided with a plurality of input channels, the map vector and the user behavior vector can be respectively used as the input of one input channel, the input map vector and the user vector can be processed through operations such as convolution kernel pooling and the like in the feature extraction model, and the information vector output by the feature extraction model can be used as the user behavior feature. For example, fig. 4 is a schematic structural diagram of a feature extraction model provided in the second embodiment of the present invention, referring to fig. 4, the feature extraction model in the second embodiment of the present invention may specifically be a multi-channel convolutional neural network model, a map vector and a user behavior vector may be regarded as input information of two input channels, a corresponding convolution kernel may be a three-dimensional matrix with a dimension of 2, and corresponding convolution operations may generally be summation after multiplication of each position of the three-dimensional matrix, where the step size is 1, and the feature extraction model may be moved laterally by 4 steps and longitudinally by 4 steps, and finally obtain a feature map with convolution kernels of 4 × 4. When the size of the atlas vector and the user phase vector input by the feature extraction model is 8 × 50 × 2, the size of convolution kernels is 1 × 50 × 2 and 2 × 50 × 2 respectively, 128 convolution kernels are provided for each size, vectors after different convolution kernel pooling operations are connected to serve as information vectors serving as user behavior features, and the vector representation dimension of the information vectors can be 256.
In an exemplary embodiment, the information recommendation method specifically includes five steps of data extraction, data analysis and training set generation, model construction and information recommendation, and the specific work involved in each step may include the following steps:
data extraction: the data acquired in the embodiment of the invention can be divided into information data and buried point data, the information data comprises information such as title and content of information to be recommended, the buried point data can comprise user information browsing records and the like, and as the acquired data can be unstructured data, related fields in the data can be extracted through regular expressions.
And (3) data analysis: and carrying out statistical analysis on related fields of the buried point data to determine the timeliness and sparseness characteristics of the information in the information recommendation process.
Generating a training set: the method comprises the steps of performing word segmentation on information data by using application software jieba, improving the word segmentation effect by adding external reserved words and stop words, and then generating a model training set and a test set required by information recommendation by adopting a negative sampling method.
Constructing a model: the model construction mainly comprises two parts of information characteristic extraction and user characteristic extraction, wherein the information characteristic extraction can be realized by using a vector direct connection method and a convolutional neural network, the convolutional neural network can improve the accuracy of user behavior characteristic determination through a financial knowledge graph, the extraction of the user behavior characteristics can be realized through a long-short term memory artificial neural network model, and further, an Attention mechanism can be added to the long-short term memory artificial model to improve the diversity of the user behavior characteristic extraction.
Furthermore, the vector direct connection method can be that word2vec method is used for word segmentation of information and Node2vec method is used for entity to convert into vectors respectively, and then all vectors are connected in the first place to be used as final expression vectors of information, so that original voice information of the information vectors can be reserved, and the accuracy of information recommendation can be improved.
Information recommendation: and selecting target information from the information to be recommended to recommend by using the determined user behavior characteristics.
EXAMPLE III
Fig. 5 is a schematic structural diagram of an information recommendation apparatus provided in the third embodiment of the present invention, which is capable of executing the information recommendation method provided in any embodiment of the present invention, and has functional modules and beneficial effects corresponding to the execution method. The device can be implemented by software and/or hardware, and specifically comprises: a graph building model 301, a user characteristic module 302 and an information recommending module 303.
The map construction model 301 is used for constructing a financial knowledge map according to entities of a preset financial data set.
A user characteristics module 302, configured to determine user behavior characteristics based on the financial knowledge-graph and the user information browsing records.
And the information recommending module 303 is configured to select target information for recommendation according to the user behavior characteristics.
According to the embodiment of the invention, the financial knowledge map is established by presetting the entity of the financial data set through the map building model, the user characteristic module determines the user behavior characteristics according to the financial knowledge map and the user information browsing record, and the information recommendation module selects the target information for recommendation based on the user behavior characteristics, so that the individuation of information recommendation is improved, the information recommendation range is enriched, and the information recommendation quality can be enhanced.
Further, on the basis of the above embodiment of the invention, the map construction model 301 includes:
and the entity identification unit is used for identifying the noun text in the preset financial data set as an entity, wherein the noun text at least comprises at least one of a place, a person name, an organization name, a company name and an industry name.
And the map establishing unit is used for reducing the dimensionality of the entity according to a preset map representation learning processing operation so as to obtain the financial knowledge map.
Further, on the basis of the above embodiment of the invention, the preset map in the map establishing unit represents that the learning process at least includes one of the following steps: graph representation learning based on TransE, graph representation learning based on TransD, graph representation learning based on TransR, Deepwalk graph representation learning and Node2vec graph representation learning.
Further, on the basis of the above embodiment of the present invention, the entity identifying unit includes:
and the similarity subunit is used for determining the similarity of each noun text through a preset text similarity model.
And the entity mapping subunit is used for mapping the noun texts with the same semantics to the same entity according to the similarity.
Further, on the basis of the above embodiment of the present invention, the user feature module 302 includes:
and the vector acquisition unit is used for acquiring the map vector corresponding to the financial knowledge map and the user behavior vector corresponding to the user information browsing record.
And the characteristic extraction unit is used for inputting the atlas vector and the user behavior vector into a characteristic extraction model to obtain user behavior characteristics, wherein the characteristic extraction model is generated by training through a historical browsing data set.
Further, on the basis of the above-described embodiment of the present invention, the feature extraction unit includes:
and the input subunit is used for respectively inputting the map vector and the user behavior vector to a feature extraction model, wherein the map vector and the user behavior vector respectively correspond to one input channel of the feature extraction model.
And the processing subunit is used for performing convolution and pooling operations on the map vector and the user behavior vector in the feature extraction model to obtain an information vector as the user behavior feature.
Further, on the basis of the above embodiment of the present invention, the information recommending module 303 includes:
and the probability unit is used for determining the recommendation probability corresponding to the information to be recommended based on the user behavior characteristics.
And the recommending unit is used for selecting the corresponding information to be recommended as target information to recommend according to the recommending probability.
Example four
Fig. 6 is a schematic structural diagram of an electronic apparatus provided in the fourth embodiment of the present invention, as shown in fig. 6, the electronic apparatus includes a processor 40, a memory 41, an input device 42, and an output device 43; the number of the processors 40 in the electronic device may be one or more, and one processor 40 is taken as an example in fig. 6; the processor 40, the memory 41, the input device 42 and the output device 43 in the electronic apparatus may be connected by a bus or other means, and the bus connection is exemplified in fig. 6.
The memory 41 is used as a computer-readable storage medium for storing software programs, computer-executable programs, and modules, such as program instructions/modules (e.g., the map building model 301, the user characteristic module 302, and the information recommending module 303 in the information recommending apparatus) corresponding to the information recommending method in the embodiment of the present invention. The processor 40 executes various functional applications and data processing of the electronic device by executing software programs, instructions and modules stored in the memory 41, so as to implement the information recommendation method.
The memory 41 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; the storage data area may store data created according to the use of the terminal, and the like. Further, the memory 41 may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid state storage device. In some examples, memory 41 may further include memory located remotely from processor 40, which may be connected to the electronic device through a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The input device 42 is operable to receive input numeric or character information and to generate key signal inputs relating to user settings and function controls of the electronic apparatus. The output device 73 may include a display device such as a display screen.
EXAMPLE five
An embodiment of the present invention further provides a storage medium containing computer-executable instructions, which when executed by a computer processor, perform an information recommendation method, including:
constructing a financial knowledge map according to entities of a preset financial data set;
determining user behavior characteristics based on the financial knowledge graph and user information browsing records;
and selecting target information for recommendation according to the user behavior characteristics.
Of course, the storage medium containing the computer-executable instructions provided by the embodiments of the present invention is not limited to the method operations described above, and may also perform related operations in the information recommendation method provided by any embodiments of the present invention.
From the above description of the embodiments, it is obvious for those skilled in the art that the present invention can be implemented by software and necessary general hardware, and certainly, can also be implemented by hardware, but the former is a better embodiment in many cases. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as a floppy disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a FLASH Memory (FLASH), a hard disk or an optical disk of a computer, and includes several instructions for enabling a computer device (which may be a personal computer, a server, or a network device) to execute the methods according to the embodiments of the present invention.
It should be noted that, in the embodiment of the information recommendation apparatus, the included units and modules are only divided according to the functional logic, but not limited to the above division, as long as the corresponding functions can be realized; in addition, specific names of the functional units are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present invention.
It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.
Claims (10)
1. An information recommendation method, the method comprising:
constructing a financial knowledge map according to entities of a preset financial data set;
determining user behavior characteristics based on the financial knowledge graph and user information browsing records;
and selecting target information for recommendation according to the user behavior characteristics.
2. The method of claim 1, wherein constructing the financial knowledge-graph from entities of the predetermined set of financial data comprises:
identifying noun texts in the preset financial data set as entities, wherein the noun texts at least comprise at least one of places, names of persons, names of institutions, names of companies and names of industries;
and reducing the dimensionality of the entity according to a preset map representation learning processing operation to obtain a financial knowledge map.
3. The method of claim 2, wherein the pre-set map representation learning process includes at least one of: graph representation learning based on TransE, graph representation learning based on TransD, graph representation learning based on TransR, Deepwalk graph representation learning and Node2vec graph representation learning.
4. The method of claim 2, wherein identifying noun text in the pre-defined set of financial data as an entity comprises:
determining the similarity of each noun text through a preset text similarity model;
and mapping noun texts with the same semantics to the same entity according to the similarity.
5. The method of claim 1, wherein determining user behavior characteristics based on the financial knowledge-graph and user information browsing records comprises:
acquiring a map vector corresponding to the financial knowledge map and a user behavior vector corresponding to a user information browsing record;
and inputting the atlas vector and the user behavior vector into a feature extraction model to obtain user behavior features, wherein the feature extraction model is generated by training of a historical browsing data set.
6. The method of claim 5, wherein inputting the atlas vector and the user behavior vector into a feature extraction model to obtain user behavior features comprises:
inputting the map vector and the user behavior vector into a feature extraction model respectively, wherein the map vector and the user behavior vector correspond to one input channel of the feature extraction model respectively;
and performing convolution and pooling operation on the atlas vector and the user behavior vector in the feature extraction model to obtain an information vector as the user behavior feature.
7. The method according to any one of claims 1-6, wherein selecting the target information for recommendation according to the user behavior characteristics comprises:
determining recommendation probability corresponding to information to be recommended based on the user behavior characteristics;
and selecting the corresponding information to be recommended as target information for recommendation according to the recommendation probability.
8. An information recommendation apparatus, comprising:
the map construction model is used for constructing a financial knowledge map according to the entities of the preset financial data set;
the user characteristic module is used for determining user behavior characteristics based on the financial knowledge map and the user information browsing record;
and the information recommendation module is used for selecting target information to recommend according to the user behavior characteristics.
9. An electronic device, characterized in that the electronic device comprises:
one or more processors;
a memory for storing one or more programs,
when executed by the one or more processors, cause the one or more processors to implement the information recommendation method of any of claims 1-7.
10. A computer-readable storage medium, on which a computer program is stored, the computer program, when executed, implementing the information recommendation method according to any one of claims 1-7.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011408670.9A CN112395506A (en) | 2020-12-04 | 2020-12-04 | Information recommendation method and device, electronic equipment and storage medium |
| PCT/CN2021/104680 WO2022116537A1 (en) | 2020-12-04 | 2021-07-06 | News recommendation method and apparatus, and electronic device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011408670.9A CN112395506A (en) | 2020-12-04 | 2020-12-04 | Information recommendation method and device, electronic equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112395506A true CN112395506A (en) | 2021-02-23 |
Family
ID=74604337
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011408670.9A Pending CN112395506A (en) | 2020-12-04 | 2020-12-04 | Information recommendation method and device, electronic equipment and storage medium |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN112395506A (en) |
| WO (1) | WO2022116537A1 (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113407729A (en) * | 2021-05-11 | 2021-09-17 | 银江股份有限公司 | Judicial-oriented personalized case recommendation method and system |
| WO2022116537A1 (en) * | 2020-12-04 | 2022-06-09 | 上海帜讯信息技术股份有限公司 | News recommendation method and apparatus, and electronic device and storage medium |
| CN116028617A (en) * | 2022-12-06 | 2023-04-28 | 腾讯科技（深圳）有限公司 | Information recommendation method, apparatus, device, readable storage medium and program product |
| CN116629984A (en) * | 2023-07-24 | 2023-08-22 | 中信证券股份有限公司 | Product information recommendation method, device, equipment and medium based on embedded model |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114969557B (en) * | 2022-07-29 | 2022-11-08 | 之江实验室 | Propaganda and education pushing method and system based on multi-source information fusion |
| CN115239442B (en) * | 2022-09-22 | 2023-01-06 | 湖南快乐通宝小额贷款有限公司 | Method and system for popularizing internet financial products and storage medium |
| CN115618108B (en) * | 2022-10-24 | 2023-05-26 | 云南中烟工业有限责任公司 | Brand recommendation method based on knowledge graph in new retail model |
| CN116089624B (en) * | 2022-11-17 | 2024-02-27 | 昆仑数智科技有限责任公司 | Knowledge graph-based data recommendation method, device and system |
| CN115757984B (en) * | 2022-11-18 | 2023-06-20 | 青海师范大学 | Fusion recommendation method based on depth migration and article collaborative filtering |
| CN115577095B (en) * | 2022-12-12 | 2023-05-02 | 南方电网科学研究院有限责任公司 | Electric power standard information recommendation method based on graph theory |
| CN115794898B (en) * | 2023-01-05 | 2023-04-25 | 中信建投证券股份有限公司 | Financial information recommendation method and device, electronic equipment and storage medium |
| CN116304303B (en) * | 2023-02-01 | 2023-09-08 | 北京三维天地科技股份有限公司 | Asset recommendation method and system based on knowledge graph |
| CN116108162B (en) * | 2023-03-02 | 2024-03-08 | 广东工业大学 | Complex text recommendation method and system based on semantic enhancement |
| CN116932488B (en) * | 2023-09-15 | 2024-01-09 | 之江实验室 | Courseware generation method, device and system based on knowledge graph and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110222198A (en) * | 2019-06-18 | 2019-09-10 | 卓尔智联(武汉)研究院有限公司 | Non-ferrous metal industry knowledge mapping construction method, electronic device and storage medium |
| CN110321482A (en) * | 2019-06-11 | 2019-10-11 | 阿里巴巴集团控股有限公司 | A kind of recommended method of information, device and equipment |
| CN111159431A (en) * | 2019-12-30 | 2020-05-15 | 深圳Tcl新技术有限公司 | Knowledge graph-based information visualization method, device, equipment and storage medium |
| CN111401055A (en) * | 2020-04-07 | 2020-07-10 | 宁波深擎信息科技有限公司 | Method and apparatus for extracting context information from financial information |
| CN111444410A (en) * | 2020-03-27 | 2020-07-24 | 民生科技有限责任公司 | Associated transaction mining and identifying method and device based on knowledge graph |
| CN111582509A (en) * | 2020-05-07 | 2020-08-25 | 南京邮电大学 | Knowledge graph representation learning and neural network based collaborative recommendation method |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109064318A (en) * | 2018-08-24 | 2018-12-21 | 苏宁消费金融有限公司 | A kind of internet financial risks monitoring system of knowledge based map |
| CN110458641B (en) * | 2019-06-28 | 2022-02-25 | 苏宁云计算有限公司 | E-commerce recommendation method and system |
| CN110795571B (en) * | 2019-10-24 | 2023-05-05 | 南宁师范大学 | Cultural travel resource recommendation method based on deep learning and knowledge graph |
| CN112395506A (en) * | 2020-12-04 | 2021-02-23 | 上海帜讯信息技术股份有限公司 | Information recommendation method and device, electronic equipment and storage medium |
-
2020
- 2020-12-04 CN CN202011408670.9A patent/CN112395506A/en active Pending
-
2021
- 2021-07-06 WO PCT/CN2021/104680 patent/WO2022116537A1/en active Application Filing
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110321482A (en) * | 2019-06-11 | 2019-10-11 | 阿里巴巴集团控股有限公司 | A kind of recommended method of information, device and equipment |
| CN110222198A (en) * | 2019-06-18 | 2019-09-10 | 卓尔智联(武汉)研究院有限公司 | Non-ferrous metal industry knowledge mapping construction method, electronic device and storage medium |
| CN111159431A (en) * | 2019-12-30 | 2020-05-15 | 深圳Tcl新技术有限公司 | Knowledge graph-based information visualization method, device, equipment and storage medium |
| CN111444410A (en) * | 2020-03-27 | 2020-07-24 | 民生科技有限责任公司 | Associated transaction mining and identifying method and device based on knowledge graph |
| CN111401055A (en) * | 2020-04-07 | 2020-07-10 | 宁波深擎信息科技有限公司 | Method and apparatus for extracting context information from financial information |
| CN111582509A (en) * | 2020-05-07 | 2020-08-25 | 南京邮电大学 | Knowledge graph representation learning and neural network based collaborative recommendation method |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2022116537A1 (en) * | 2020-12-04 | 2022-06-09 | 上海帜讯信息技术股份有限公司 | News recommendation method and apparatus, and electronic device and storage medium |
| CN113407729A (en) * | 2021-05-11 | 2021-09-17 | 银江股份有限公司 | Judicial-oriented personalized case recommendation method and system |
| CN113407729B (en) * | 2021-05-11 | 2022-06-24 | 银江技术股份有限公司 | Judicial-oriented personalized case recommendation method and system |
| CN116028617A (en) * | 2022-12-06 | 2023-04-28 | 腾讯科技（深圳）有限公司 | Information recommendation method, apparatus, device, readable storage medium and program product |
| CN116028617B (en) * | 2022-12-06 | 2024-02-27 | 腾讯科技（深圳）有限公司 | Information recommendation method, apparatus, device, readable storage medium and program product |
| CN116629984A (en) * | 2023-07-24 | 2023-08-22 | 中信证券股份有限公司 | Product information recommendation method, device, equipment and medium based on embedded model |
| CN116629984B (en) * | 2023-07-24 | 2024-02-06 | 中信证券股份有限公司 | Product information recommendation method, device, equipment and medium based on embedded model |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2022116537A1 (en) | 2022-06-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112395506A (en) | Information recommendation method and device, electronic equipment and storage medium | |
| CN111753060B (en) | Information retrieval method, apparatus, device and computer readable storage medium | |
| CN111444428B (en) | Information recommendation method and device based on artificial intelligence, electronic equipment and storage medium | |
| CN110069709B (en) | Intention recognition method, device, computer readable medium and electronic equipment | |
| CN112131350B (en) | Text label determining method, device, terminal and readable storage medium | |
| CN109471944B (en) | Training method and device of text classification model and readable storage medium | |
| CN111324771B (en) | Video tag determination method and device, electronic equipment and storage medium | |
| CN110619051B (en) | Question sentence classification method, device, electronic equipment and storage medium | |
| WO2005022413A1 (en) | Method and apparatus for automatic file clustering into a data-driven, user-specific taxonomy | |
| CN111539197A (en) | Text matching method and device, computer system and readable storage medium | |
| WO2023108993A1 (en) | Product recommendation method, apparatus and device based on deep clustering algorithm, and medium | |
| CN113254711B (en) | Interactive image display method and device, computer equipment and storage medium | |
| CN111783903B (en) | Text processing method, text model processing method and device and computer equipment | |
| CN111666766A (en) | Data processing method, device and equipment | |
| CN113806588A (en) | Method and device for searching video | |
| CN114416998A (en) | Text label identification method and device, electronic equipment and storage medium | |
| CN115795030A (en) | Text classification method and device, computer equipment and storage medium | |
| CN113282754A (en) | Public opinion detection method, device, equipment and storage medium for news events | |
| CN114090793A (en) | Information extraction method and device, electronic equipment, computer readable medium and product | |
| CN116882414B (en) | Automatic comment generation method and related device based on large-scale language model | |
| CN114328800A (en) | Text processing method and device, electronic equipment and computer readable storage medium | |
| CN114547257B (en) | Class matching method and device, computer equipment and storage medium | |
| CN117216617A (en) | Text classification model training method, device, computer equipment and storage medium | |
| CN115129890A (en) | Feedback data map generation method and generation device, question answering device and refrigerator | |
| CN114676237A (en) | Sentence similarity determining method and device, computer equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |