WO2022116537A1 - Procédé et appareil de recommandation de nouvelles, ainsi que dispositif électronique et support de stockage - Google Patents

Procédé et appareil de recommandation de nouvelles, ainsi que dispositif électronique et support de stockage Download PDF

Info

Publication number
WO2022116537A1
WO2022116537A1 PCT/CN2021/104680 CN2021104680W WO2022116537A1 WO 2022116537 A1 WO2022116537 A1 WO 2022116537A1 CN 2021104680 W CN2021104680 W CN 2021104680W WO 2022116537 A1 WO2022116537 A1 WO 2022116537A1
Authority
WO
WIPO (PCT)
Prior art keywords
vector
information
entity
user behavior
recommendation
Prior art date
Application number
PCT/CN2021/104680
Other languages
English (en)
Chinese (zh)
Inventor
沈浩
黄海量
Original Assignee
上海帜讯信息技术股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 上海帜讯信息技术股份有限公司 filed Critical 上海帜讯信息技术股份有限公司
Publication of WO2022116537A1 publication Critical patent/WO2022116537A1/fr

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

Definitions

  • Embodiments of the present invention relate to the technical field of computer applications, and in particular, to an information recommendation method, apparatus, electronic device, and storage medium.
  • the invention provides an information recommendation method, device, electronic equipment and storage medium, which can increase external knowledge through knowledge graph, improve the richness of information recommendation, enhance the quality of information recommendation, and improve the user experience.
  • an embodiment of the present invention provides an information recommendation method, and the method includes:
  • target information is selected for recommendation.
  • an embodiment of the present invention further provides an information recommendation device, the device comprising:
  • the graph construction model is used to construct a financial knowledge graph based on the entities of the preset financial data set
  • a user feature module configured to determine user behavior features based on the financial knowledge graph and user information browsing records
  • the information recommendation module is used for selecting target information for recommendation according to the user behavior characteristics.
  • an embodiment of the present invention further provides an electronic device, the electronic device comprising:
  • processors one or more processors
  • memory for storing one or more programs
  • the one or more processors 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.
  • an embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, implements the method according to any one of the embodiments of the present invention.
  • a financial knowledge graph is established by an entity of a preset financial data set, user behavior characteristics are determined through the financial knowledge graph and user information browsing records, and target information is selected based on the user behavior characteristics for recommendation, which improves the personalization of information recommendation.
  • the range of information recommendation is enriched and the quality of information recommendation can be enhanced.
  • Embodiment 1 is a flowchart of an information recommendation method in Embodiment 1 of the present invention.
  • Embodiment 2 is a flowchart of an information recommendation method in Embodiment 2 of the present invention.
  • Embodiment 3 is an example diagram of a financial knowledge graph provided in Embodiment 2 of the present invention.
  • FIG. 4 is a schematic structural diagram of a feature extraction model provided in Embodiment 2 of the present invention.
  • FIG. 5 is a schematic structural diagram of an information recommendation device provided in Embodiment 3 of the present invention.
  • FIG. 6 is a schematic structural diagram of an electronic device provided in Embodiment 4 of the present invention.
  • FIG. 1 is a flowchart of an information recommendation method in Embodiment 1 of the present invention. This embodiment can be applied to the situation of personalized recommendation of information information.
  • the method can be executed by an information recommendation device, and the device can use hardware and/or software.
  • 1, the method provided by the embodiment of the present invention may generally include the following steps:
  • Step 110 Build a financial knowledge graph according to the entities of the preset financial data set.
  • the preset financial data set may be a pre-collected data set, the preset financial data set may include company information and industry information, etc., and the preset financial data set may be obtained from Wind. All listed companies in Shenzhen and Shenzhen have industry classification information.
  • the financial knowledge graph can be information for mining and displaying knowledge in the financial field. Each edge in the financial knowledge graph exists in the form of an entity-relation triple (h, r, t), where h, r, t are respectively Represents financial entities, relationships, tail entities, etc.
  • 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 acquired.
  • a knowledge graph can be constructed according to each entity and the relationship between each entity. Exemplarily, 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 between the vertices.
  • Step 120 Determine user behavior characteristics based on the financial knowledge graph and user information browsing records.
  • 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 the title, content and browsing time browsed by the user.
  • the user behavior feature may be a habit feature reflecting the user's browsing financial information, and the user behavior feature may specifically be a vector, and the vector may include financial information for browsing or disliking browsing.
  • feature extraction can be performed on the user information browsing records based on the financial knowledge graph, and the extracted features can be used as user behavior characteristics.
  • the record is searched for the corresponding entity in the financial knowledge graph, and the financial information corresponding to the entity can be used as the user behavior feature.
  • the financial knowledge graph and user browsing records can also be input into the neural network model for feature extraction, and the extraction results can be used as user behavior characteristics. .
  • Step 130 Select target information for recommendation according to the user behavior characteristics.
  • the target information may be information corresponding to the user behavior characteristics, the target information may be selected and generated from the information set to be recommended, and the selection principle may be the user behavior characteristics.
  • one or more pieces of information may be selected from the information set to be recommended as target information for recommendation based on the information content included in the user behavior characteristics.
  • the recommendation probability corresponding to the information information in the information set to be recommended can also be predicted according to the user behavior characteristics, and one or more information information can be selected as the target information for recommendation according to the recommendation probability in descending order.
  • a financial knowledge graph is established by extracting entities of a preset financial data set, user behavior characteristics are obtained by analyzing user information browsing records based on the financial knowledge graph, and target information is selected based on the user behavior characteristics for recommendation, thereby improving information efficiency.
  • the personalized recommendation enriches the scope of information recommendation and enhances the quality of information recommendation.
  • FIG. 2 is a flow chart of the information recommendation method in the second embodiment of the present invention.
  • the embodiment of the present invention is a specific embodiment based on the above-mentioned embodiment of the present invention. Referring to FIG. 2 , the method provided by the embodiment of the present invention specifically includes the following steps:
  • Step 210 Identify the noun text in the preset financial data set as an entity, and the noun text at least includes at least one of a place, a person's name, an institution name, a company name, and an industry name.
  • the noun text can be a vocabulary with actual meaning, which can include place name, person name, institution name, company name and industry name, etc.
  • the entity can be the constituent unit of the financial knowledge graph, and the entity and noun text can have a corresponding relationship, one or more
  • the noun text can correspond to an entity.
  • iFLYTEK and NavInfo can correspond to the same entity, which can represent the technological innovation sector in the financial knowledge graph.
  • different word segmentations can be obtained by performing word segmentation operations on the text in the preset financial data set. Since word segmentations may contain text words that do not have actual meanings, the obtained word segmentations can be directly identified and obtained by using the thesaurus to a noun text with actual meaning as an entity. Further, due to the diversity and ambiguity of natural semantic expressions in the noun texts identified from the preset financial data set, there may be multiple noun texts with different expressions for the same entity.
  • the semantic fuzzy machine learning model can be used to calculate each noun. Text similarity can map noun texts with the same semantics to the same entity.
  • Step 220 according to the preset graph representation learning processing operation, reduce the dimension of the entity to obtain the 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 save the intrinsic structural information of the cause.
  • the dimension of the entity can be reduced according to the preset graph representation learning processing operation to be embedded in the low-dimensional vector space, and the low-dimensional vector space embedded with the entity can be used as a financial knowledge graph, so that a continuous form is formed after dimensionality reduction.
  • Spatial vector, reducing the data processing volume of the subsequent information recommendation process, each edge in the financial knowledge graph can be represented by an entity-relation triple (h, r, t), h is the vector of the head entity, r is the head entity and A vector of relationships between tail entities, where t is a vector of tail entities.
  • the preset graph representation learning processing operations may include two types of processing operations based on Trans and based on paths.
  • the financial knowledge graph can include two types of entities: companies and industries, as well as a type of relationship in which the company belongs to the industry.
  • the Trans-based method is more suitable for the representation of graphs with many types of relationships, while the path-based method is more suitable for relatively close distances.
  • Graph representation of entities having similar properties, in this embodiment of the present invention, a path-based graph can be used to represent processing operations.
  • the dimension of the entity vector is set to 50, and the obtained entity vector is dimensionally reduced to form a financial knowledge graph as shown in Figure 3.
  • Each entity in different industries is represented by icons of different shapes, and it can be seen that the same
  • the entities of the industry are all clustered in the same cluster, and the entities of different industries can also be clearly distinguished.
  • the cluster in the lower left corner of the figure is companies in the machinery and equipment industry
  • the cluster in the middle is companies in the pharmaceutical and biological industries.
  • Step 230 Obtain a graph vector corresponding to the financial knowledge graph and a user behavior vector corresponding to the user information browsing record.
  • the financial knowledge graph can be represented in the form of a vector, for example, the entity distribution in the financial knowledge graph can be used as an element of the vector.
  • the information title or information label in the user information browsing record can be extracted, and the obtained title or label can be used as a component element in the user behavior vector.
  • Step 240 inputting the atlas vector and the user behavior vector into the feature extraction model to obtain the user behavior feature, wherein the feature extraction model is generated by training the historical browsing data set.
  • the feature extraction model may be a model for extracting features of users' reading habits, specifically a convolutional neural network model, and the feature extraction model may include one or more input channels.
  • the user behavior feature may specifically be the feature of the user reading information, such as the type of information that you like to browse the most or the type of information that you don’t like to browse the most.
  • the data scale of the browsing data set can meet the threshold condition. The larger the data scale of the historical browsing data set, the higher the accuracy of the feature extraction model generated by the training of historical browsing data in extracting user behavior features.
  • the historical browsing data set can be used as a positive sample of the training set, and a data set can be randomly generated as a negative sample, and the feature extraction model can be trained by using the positive samples and negative samples, until the output of the extraction model is extracted.
  • the result meets the requirements.
  • the acquired atlas vector and user behavior vector may be input into the feature extraction model, the atlas vector and user behavior vector may be fused as an input vector and input into the feature extraction model, or the atlas vector may be input into the feature extraction model.
  • the user behavior vector are respectively input into the feature extraction model as a vector, and the user features can be obtained by processing the graph vector and the user behavior vector through the feature extraction model.
  • Step 250 Determine a 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.
  • the user behavior characteristics can be used to determine the recommendation probability of each recommended information in the information to be recommended.
  • the higher the recommendation probability the more likely the corresponding recommended information to be read by the user.
  • the information reading history and information of a large number of users can be used.
  • the information to be recommended trains a long and short-term memory network model, which is used to determine the recommendation probability corresponding to the information to be recommended, and the user behavior characteristics and the information to be recommended can be input into the long and short-term memory network model to obtain the corresponding
  • the information to be recommended can include multiple pieces 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 the form of a vector to obtain the recommendation probability corresponding to each information to be recommended.
  • Step 260 Select corresponding information to be recommended as target information for recommendation according to the recommendation probability.
  • the corresponding information to be recommended may be selected as target information to be recommended to the user according to the recommendation probability, and the information to be recommended with a larger value of the recommendation probability may be selected first as the target information.
  • 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.
  • the present invention by identifying the noun text in the preset financial data set as the entity, representing the learning operation according to the preset graph, reducing the dimension of the entity and constructing the financial knowledge graph, and obtaining the graph vector corresponding to the financial knowledge graph and the corresponding user information browsing records.
  • the user behavior vector is input into the feature extraction model to obtain user behavior characteristics, the recommendation probability of the information to be recommended is determined based on the user behavior characteristics, and the target information is obtained according to the recommendation probability for recommendation, which improves the personalization degree of user information recommendation and enriches information recommendation. Scope to enhance the user's experience of reading information.
  • the preset graph representation learning process includes at least one of the following: TransE-based graph representation learning, TransD-based graph representation learning, TransR-based graph representation learning, DeepWalk graph representation Learning, Node2vec graph representation learning.
  • TransE-based graph representation learning can target an edge of a financial knowledge graph corresponding to an entity-relation triple (h, r, t), where h, r, and t represent head entities, relationships, and relationships, respectively.
  • the tail entity assuming that (h, r, t) satisfies h+r ⁇ t, where h, r, t represent the vector representation of the head entity, the relationship, and the tail entity respectively, that is, the one that satisfies the (h, r, t) relationship
  • the TransD-based graph representation learning assumes that each different type of entity relationship has a mappable hyperplane, and each entity is first mapped to the hyperplane of the corresponding relationship by the following formula:
  • W r is the normal vector of the r relational hyperplane
  • T is the matrix transpose
  • h ⁇ is the vector of the mapped head entity
  • t ⁇ is the vector of the mapped tail entity.
  • TransR-based graph representation learning assumes that there is a relation space for each different type of entity relation, and entities are mapped to the corresponding relation space through the mapping matrix M r :
  • the mapped vector also satisfies the score function of TransE, namely:
  • DeepWalk graph representation learning can use the simplest random walk method to generate a sequence of nodes, that is, the method of selecting the next node is uniformly distributed randomly, and the nodes in the sequence can be regarded as words in the text, using Skip- Gram model to train a vector of nodes.
  • Node2vec graph representation learning can be based on changing the way DeepWalk random walk sequence is generated. By introducing two parameters p and q, breadth-first search and depth-first search are introduced into the generation process of random walk sequence.
  • the identifying the noun text in the preset financial data set as an entity includes: determining the similarity of each of the noun texts through a preset text similarity model; Degree maps noun texts with the same semantics to the same entities.
  • the preset text similarity model may be a deep learning model for determining the similarity between noun texts, and the model may be generated through training of massive text vocabulary.
  • the noun text identified in the preset financial data set can be input into the preset text similarity model, and the output result can be obtained.
  • the output result can be a result vector, each element can correspond to a noun text, and the numerical value in the result vector can be calculated. Noun texts that are the same or whose difference is less than a fixed threshold are mapped to the same entity as the same type of text.
  • inputting the atlas vector and the user behavior vector into the feature extraction model to obtain user behavior features includes:
  • the atlas vector and the user behavior vector are respectively input into the feature extraction model, wherein the atlas vector and the user behavior vector respectively correspond to an input channel of the feature extraction model; Convolution and pooling operations are performed on the graph vector and the user behavior vector to obtain the information vector as the user behavior feature.
  • the feature extraction model may have multiple input channels, the atlas vector and the user behavior vector may be used as the input of one input channel respectively, and the input atlas vector and User vector, the information vector output by the feature extraction model can be used as the user behavior feature.
  • FIG. 4 is a schematic structural diagram of a feature extraction model provided in Embodiment 2 of the present invention.
  • the feature extraction model in this embodiment of the present invention may specifically be a multi-channel convolutional neural network model, which may be The map vector and the user behavior vector are regarded as the input information of the two input channels respectively.
  • the corresponding convolution kernel can be a three-dimensional matrix with a dimension of 2.
  • the corresponding convolution operation can be generalized by multiplying each position of the three-dimensional matrix. Summation, with a step size of 1, can move 4 steps horizontally and 4 steps vertically, and finally obtain a feature map of a 4*4 convolution kernel.
  • the size of the map vector and the user phase vector input by the feature extraction model is 8*50*2
  • the size of the convolution kernel is 1*50*2 and 2*50*2 respectively
  • the convolution kernel of each size has 128
  • Each of the vectors after the pooling operation of different convolution kernels is connected and can be used as an information vector as a user behavior feature, wherein the vector representation dimension of the information vector can be 256.
  • the information recommendation method may specifically include five steps of data extraction, data analysis and training set generation, model building and information recommendation, and the specific work involved in each step may include the following:
  • the data obtained in the embodiment of the present invention can be divided into information data and buried point data.
  • the information data includes information such as the title and content of the information to be recommended, and the buried point data can include user information browsing records, etc.
  • the data can be unstructured data, and relevant fields in the data can be extracted through regular expressions.
  • Data analysis Statistical analysis is performed on the relevant fields of the embedded point data to determine the timeliness and sparsity characteristics of the information in the information recommendation process.
  • Training set generation Use the application software jieba to segment the information data, and improve the effect of word segmentation by adding external reservations and stop words, and then use the negative sampling method to generate the model training set and test set required for information recommendation.
  • Model construction mainly includes two parts: information feature extraction and user feature extraction.
  • Information feature extraction can be realized by vector direct connection method and convolutional neural network.
  • convolutional neural network can improve user behavior feature determination through financial knowledge graph
  • the accuracy of user behavior features can be achieved through the long short-term memory artificial neural network model.
  • the Attention mechanism can be added to the long short-term memory artificial model to improve the diversity of user behavior feature extraction.
  • the vector direct connection method may be to use the word2vec method for the word segmentation of the information and the Node2vec method for the entities to convert them into vectors respectively, and then connect the first positions of all vectors as the final representation vector of the information, which can retain the original voice information of the information vector. , which helps to improve the accuracy of information recommendation.
  • Information recommendation You can use the determined user behavior characteristics to select target information from the information to be recommended for recommendation.
  • Embodiment 5 is a schematic structural diagram of an information recommendation apparatus provided in Embodiment 3 of the present invention, which can execute the information recommendation method provided by any of the embodiments of the present invention, and has functional modules and beneficial effects corresponding to the execution method.
  • the apparatus may be implemented by software and/or hardware, and specifically includes: a graph construction model 301 , a user feature module 302 and an information recommendation module 303 .
  • the graph construction model 301 is used to construct a financial knowledge graph according to the entities of the preset financial data set.
  • the user feature module 302 is configured to determine user behavior features based on the financial knowledge graph and user information browsing records.
  • the information recommendation module 303 is configured to select target information for recommendation according to the user behavior characteristics.
  • the entity of the financial data set is preset by the graph construction model to establish a financial knowledge graph
  • the user feature module determines user behavior characteristics according to the financial knowledge graph and user information browsing records
  • the information recommendation module selects target information based on the user behavior characteristics for recommendation. , which improves the personalization of information recommendation, enriches the scope of information recommendation, and enhances the quality of information recommendation.
  • the graph construction model 301 includes:
  • the entity identification unit is used to identify the noun text in the preset financial data set as an entity, and the noun text at least includes at least one of a place, a person's name, an institution name, a company name, and an industry name.
  • a graph establishment unit configured to reduce the dimension of the entity according to the preset graph representation learning processing operation to obtain a financial knowledge graph.
  • the preset graph representation learning processing in the graph establishment unit includes at least one of the following: TransE-based graph representation learning, TransD-based graph representation learning, and TransR-based graph representation learning. , DeepWalk graph representation learning, Node2vec graph representation learning.
  • the entity identification unit includes:
  • the similarity subunit is used to determine the similarity of each of the noun texts through a preset text similarity model.
  • the entity mapping subunit is used to map the noun texts with the same semantics to the same entity according to the similarity.
  • the user feature module 302 includes:
  • a vector obtaining unit configured to obtain the graph vector corresponding to the financial knowledge graph and the user behavior vector corresponding to the user information browsing record.
  • a feature extraction unit configured to input 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 a historical browsing data set.
  • the feature extraction unit includes:
  • the input subunit is configured to input the graph vector and the user behavior vector into the feature extraction model respectively, wherein the graph vector and the user behavior vector respectively correspond to an input channel of the feature extraction model.
  • a processing subunit configured to perform convolution and pooling operations on the graph vector and the user behavior vector in the feature extraction model to obtain an information vector as a user behavior feature.
  • the information recommendation module 303 includes:
  • a probability unit configured to determine a recommendation probability corresponding to the information to be recommended based on the user behavior feature.
  • a recommending unit configured to select corresponding information to be recommended as target information for recommendation according to the recommendation probability.
  • FIG. 6 is a schematic structural diagram of an electronic device provided in Embodiment 4 of the present invention.
  • the electronic device includes a processor 40, a memory 41, an input device 42 and an output device 43; the processor 40 in the electronic device The number can be one or more, and a processor 40 is taken as an example in FIG. 6; the processor 40, memory 41, input device 42 and output device 43 in the electronic device can be connected through a bus or in other ways. Connecting via a bus is an example.
  • the memory 41 can be used to store software programs, computer-executable programs, and modules, such as program instructions/modules corresponding to the information recommendation method in the embodiment of the present invention (for example, the map construction in the information recommendation device). model 301, user feature module 302 and information recommendation module 303).
  • the processor 40 executes various functional applications and data processing of the electronic device by running the software programs, instructions and modules stored in the memory 41 , that is, to implement the above-mentioned 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.
  • 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.
  • 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, an intranet, a local area network, a mobile communication network, and combinations thereof.
  • the input device 42 may be used to receive input numerical or character information, and to generate key signal input related to user settings and function control of the electronic device.
  • the output device 73 may include a display device such as a display screen.
  • Embodiment 5 of the present invention also provides a storage medium containing computer-executable instructions, where the computer-executable instructions are used to execute an information recommendation method when executed by a computer processor, and the method includes:
  • target information is selected for recommendation.
  • a storage medium containing computer-executable instructions provided by an embodiment of the present invention, the computer-executable instructions of which are not limited to the above-mentioned method operations, and can also execute any of the information recommendation methods provided by any embodiment of the present invention. related operations.
  • the present invention can be realized by software and necessary general-purpose hardware, and of course can also be realized by hardware, but in many cases the former is a better embodiment .
  • the technical solutions of the present invention can be embodied in the form of software products in essence or the parts that make contributions to the prior art, and the computer software products can be stored in a computer-readable storage medium, such as a floppy disk of a computer , read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), flash memory (FLASH), hard disk or optical disk, etc., including several instructions to make a computer device (which can be a personal computer , server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.
  • a computer-readable storage medium such as a floppy disk of a computer , read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), flash memory (FLASH), hard disk or optical disk, etc.
  • the units and modules included are only divided according to functional logic, but are not limited to the above-mentioned division, as long as the corresponding functions can be realized;
  • the specific names of the functional units are only for the convenience of distinguishing from each other, and are not used to limit the protection scope of the present invention.

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

L'invention divulgue un procédé et un appareil de recommandation de nouvelles, ainsi qu'un dispositif électronique et un support de stockage. Le procédé consiste : à construire un graphe de connaissances financières selon une entité d'un ensemble de données financières prédéfini ; à déterminer une caractéristique de comportement d'utilisateur sur la base du graphique de connaissances financières et d'un enregistrement de navigation d'informations d'utilisateur ; et, selon la caractéristique de comportement d'utilisateur, à sélectionner des nouvelles cibles pour une recommandation. Dans les modes de réalisation de la présente invention, au moyen d'un graphe de connaissances financières, la connaissance du contexte de nouvelles est identifiée et la représentation sémantique des nouvelles est enrichie, ce qui fait en sorte que la précision d'une recommandation personnalisée de nouvelles pour des utilisateurs est améliorée, et que la qualité de recommandation de nouvelles peut être améliorée.
PCT/CN2021/104680 2020-12-04 2021-07-06 Procédé et appareil de recommandation de nouvelles, ainsi que dispositif électronique et support de stockage WO2022116537A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202011408670.9A CN112395506A (zh) 2020-12-04 2020-12-04 一种资讯推荐方法、装置、电子设备和存储介质
CN202011408670.9 2020-12-04

Publications (1)

Publication Number Publication Date
WO2022116537A1 true WO2022116537A1 (fr) 2022-06-09

Family

ID=74604337

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/104680 WO2022116537A1 (fr) 2020-12-04 2021-07-06 Procédé et appareil de recommandation de nouvelles, ainsi que dispositif électronique et support de stockage

Country Status (2)

Country Link
CN (1) CN112395506A (fr)
WO (1) WO2022116537A1 (fr)

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114969557A (zh) * 2022-07-29 2022-08-30 之江实验室 一种基于多来源信息融合的宣教推送方法和系统
CN115239442A (zh) * 2022-09-22 2022-10-25 湖南快乐通宝小额贷款有限公司 一种互联网金融产品推广的方法及系统和存储介质
CN115577095A (zh) * 2022-12-12 2023-01-06 南方电网科学研究院有限责任公司 一种基于图论的电力标准信息推荐方法
CN115618108A (zh) * 2022-10-24 2023-01-17 云南中烟工业有限责任公司 新零售模式下基于知识图谱的卷烟品牌推荐方法
CN115757984A (zh) * 2022-11-18 2023-03-07 青海师范大学 基于深度游走和物品协同过滤的融合推荐方法
CN115794898A (zh) * 2023-01-05 2023-03-14 中信建投证券股份有限公司 一种金融资讯推荐方法、装置、电子设备及存储介质
CN116089624A (zh) * 2022-11-17 2023-05-09 昆仑数智科技有限责任公司 基于知识图谱的数据推荐方法、装置和系统
CN116108162A (zh) * 2023-03-02 2023-05-12 广东工业大学 一种基于语义增强的复杂文本推荐方法及系统
CN116304303A (zh) * 2023-02-01 2023-06-23 北京三维天地科技股份有限公司 一种基于知识图谱的资产推荐方法及系统
CN116932488A (zh) * 2023-09-15 2023-10-24 之江实验室 基于知识图谱的课件生成方法、装置、系统和存储介质

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112395506A (zh) * 2020-12-04 2021-02-23 上海帜讯信息技术股份有限公司 一种资讯推荐方法、装置、电子设备和存储介质
CN113407729B (zh) * 2021-05-11 2022-06-24 银江技术股份有限公司 一种面向司法的个性化案例推荐方法及系统
CN116028617B (zh) * 2022-12-06 2024-02-27 腾讯科技（深圳）有限公司 资讯推荐方法、装置、设备、可读存储介质及程序产品
CN116629984B (zh) * 2023-07-24 2024-02-06 中信证券股份有限公司 基于嵌入模型的产品信息推荐方法、装置、设备和介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109064318A (zh) * 2018-08-24 2018-12-21 苏宁消费金融有限公司 一种基于知识图谱的互联网金融风险监测系统
CN110321482A (zh) * 2019-06-11 2019-10-11 阿里巴巴集团控股有限公司 一种信息的推荐方法、装置及设备
CN110458641A (zh) * 2019-06-28 2019-11-15 苏宁云计算有限公司 一种电商推荐方法及系统
CN110795571A (zh) * 2019-10-24 2020-02-14 南宁师范大学 基于深度学习和知识图谱的文化旅游资源推荐方法
CN112395506A (zh) * 2020-12-04 2021-02-23 上海帜讯信息技术股份有限公司 一种资讯推荐方法、装置、电子设备和存储介质

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110222198A (zh) * 2019-06-18 2019-09-10 卓尔智联(武汉)研究院有限公司 有色金属行业知识图谱构建方法、电子装置及存储介质
CN111159431A (zh) * 2019-12-30 2020-05-15 深圳Tcl新技术有限公司 基于知识图谱的信息可视化方法、装置、设备及存储介质
CN111444410B (zh) * 2020-03-27 2023-09-15 民生科技有限责任公司 一种基于知识图谱的关联交易挖掘识别方法及装置
CN111401055B (zh) * 2020-04-07 2023-04-18 宁波深擎信息科技有限公司 从金融资讯提取脉络信息的方法和装置
CN111582509B (zh) * 2020-05-07 2022-09-02 南京邮电大学 一种基于知识图谱表示学习和神经网络的协同推荐方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109064318A (zh) * 2018-08-24 2018-12-21 苏宁消费金融有限公司 一种基于知识图谱的互联网金融风险监测系统
CN110321482A (zh) * 2019-06-11 2019-10-11 阿里巴巴集团控股有限公司 一种信息的推荐方法、装置及设备
CN110458641A (zh) * 2019-06-28 2019-11-15 苏宁云计算有限公司 一种电商推荐方法及系统
CN110795571A (zh) * 2019-10-24 2020-02-14 南宁师范大学 基于深度学习和知识图谱的文化旅游资源推荐方法
CN112395506A (zh) * 2020-12-04 2021-02-23 上海帜讯信息技术股份有限公司 一种资讯推荐方法、装置、电子设备和存储介质

Cited By (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114969557A (zh) * 2022-07-29 2022-08-30 之江实验室 一种基于多来源信息融合的宣教推送方法和系统
CN115239442A (zh) * 2022-09-22 2022-10-25 湖南快乐通宝小额贷款有限公司 一种互联网金融产品推广的方法及系统和存储介质
CN115618108A (zh) * 2022-10-24 2023-01-17 云南中烟工业有限责任公司 新零售模式下基于知识图谱的卷烟品牌推荐方法
CN115618108B (zh) * 2022-10-24 2023-05-26 云南中烟工业有限责任公司 新零售模式下基于知识图谱的品牌推荐方法
CN116089624A (zh) * 2022-11-17 2023-05-09 昆仑数智科技有限责任公司 基于知识图谱的数据推荐方法、装置和系统
CN116089624B (zh) * 2022-11-17 2024-02-27 昆仑数智科技有限责任公司 基于知识图谱的数据推荐方法、装置和系统
CN115757984A (zh) * 2022-11-18 2023-03-07 青海师范大学 基于深度游走和物品协同过滤的融合推荐方法
CN115757984B (zh) * 2022-11-18 2023-06-20 青海师范大学 基于深度游走和物品协同过滤的融合推荐方法
CN115577095A (zh) * 2022-12-12 2023-01-06 南方电网科学研究院有限责任公司 一种基于图论的电力标准信息推荐方法
CN115794898A (zh) * 2023-01-05 2023-03-14 中信建投证券股份有限公司 一种金融资讯推荐方法、装置、电子设备及存储介质
CN116304303A (zh) * 2023-02-01 2023-06-23 北京三维天地科技股份有限公司 一种基于知识图谱的资产推荐方法及系统
CN116304303B (zh) * 2023-02-01 2023-09-08 北京三维天地科技股份有限公司 一种基于知识图谱的资产推荐方法及系统
CN116108162A (zh) * 2023-03-02 2023-05-12 广东工业大学 一种基于语义增强的复杂文本推荐方法及系统
CN116108162B (zh) * 2023-03-02 2024-03-08 广东工业大学 一种基于语义增强的复杂文本推荐方法及系统
CN116932488A (zh) * 2023-09-15 2023-10-24 之江实验室 基于知识图谱的课件生成方法、装置、系统和存储介质
CN116932488B (zh) * 2023-09-15 2024-01-09 之江实验室 基于知识图谱的课件生成方法、装置、系统和存储介质

Also Published As

Publication number Publication date
CN112395506A (zh) 2021-02-23

Similar Documents

Publication Publication Date Title
WO2022116537A1 (fr) Procédé et appareil de recommandation de nouvelles, ainsi que dispositif électronique et support de stockage
CN111753060B (zh) 信息检索方法、装置、设备及计算机可读存储介质
WO2020108608A1 (fr) Procédé de traitement de résultat de recherche, dispositif, terminal, dispositif électronique et support de stockage
CN111680173A (zh) 统一检索跨媒体信息的cmr模型
CN106202256B (zh) 基于语义传播及混合多示例学习的Web图像检索方法
WO2018072071A1 (fr) Système et procédé de construction de carte de connaissances
US20050044487A1 (en) Method and apparatus for automatic file clustering into a data-driven, user-specific taxonomy
CN111539197A (zh) 文本匹配方法和装置以及计算机系统和可读存储介质
CN109471944A (zh) 文本分类模型的训练方法、装置及可读存储介质
CN111783903B (zh) 文本处理方法、文本模型的处理方法及装置、计算机设备
US11928418B2 (en) Text style and emphasis suggestions
Du et al. Classifying natural-language spatial relation terms with random forest algorithm
CN116848490A (zh) 使用模型相交进行文档分析
CN115795030A (zh) 文本分类方法、装置、计算机设备和存储介质
CN114997288A (zh) 一种设计资源关联方法
CN116975271A (zh) 文本相关性的确定方法、装置、计算机设备和存储介质
CN113283432A (zh) 图像识别、文字排序方法及设备
CN114090793A (zh) 信息抽取方法、装置、电子设备及计算机可读介质及产品
WO2024114681A1 (fr) Procédé et appareil d'affichage de résultats de recherche, et dispositif informatique et support de stockage
CN114254116A (zh) 文献资料文本分类方法、分类模型构建方法和分类装置
CN116882414B (zh) 基于大规模语言模型的评语自动生成方法及相关装置
CN113988057A (zh) 基于概念抽取的标题生成方法、装置、设备及介质
WO2023246849A1 (fr) Procédé de génération de graphe de données de rétroaction et réfrigérateur
CN114547257B (zh) 类案匹配方法、装置、计算机设备及存储介质
CN116521892A (zh) 知识图谱的应用方法、装置、电子设备、介质和程序产品

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21899578

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21899578

Country of ref document: EP

Kind code of ref document: A1