Information push method, device, electronic device and computer readable medium

Abstract

The present application discloses an information push method, device, electronic device and computer-readable medium, and relates to the technical field of big data analysis and mining. The method includes: classifying the labels of each transaction subject, thereby constructing a label system for each transaction subject; obtaining the source data corresponding to the label system, extracting the high-level features and low-level features of the source data and the label system, and then fusing the high-level features and low-level features to obtain fusion features, and constructing user portraits of each transaction subject based on the fusion features; matching each user portrait with the candidate push information for similarity, so as to determine the target push information corresponding to each user portrait from the candidate push information, and push the target push information to the transaction subject corresponding to the corresponding user portrait. Solve the technical problem that the information communication between suppliers, distributors and consumers is not smooth, and accurate agricultural trade information cannot be obtained.

Description

Technical Field

[0001] The present invention relates to the field of big data analysis and mining technology, and in particular to an information push method, device, electronic device and computer-readable medium. Background Art

[0002] As an important link in the field of agricultural circulation, the scale and efficiency of the agricultural wholesale market are constantly developing, and the market's functional formats are constantly enriching and improving. Agricultural wholesale markets in many parts of the country have gradually introduced the construction of smart agricultural wholesale information technology, mainly using information technologies such as cloud computing, big data, the Internet of Things, and mobile APPs to achieve digital transformation of the core functions of agricultural wholesale markets such as transactions, settlements, and related functions such as logistics distribution, financial services, and other links.

[0003] In the process of implementing the present invention, the inventors found that there are at least the following problems in the prior art:

[0004] The information communication between suppliers, distributors and consumers in the agricultural wholesale supply and marketing chain mostly relies on traditional offline methods or simple information release platforms, which cannot achieve accurate recommendation of agricultural trade information. At present, the information communication between suppliers, distributors and consumers is not smooth, and it is impossible to obtain accurate agricultural trade information. Summary of the invention

[0005] In view of this, the embodiments of the present invention provide an information push method, device, electronic device and computer-readable medium to solve the technical problem that information communication between suppliers, distributors and consumers is poor and accurate agricultural trade information cannot be obtained.

[0006] To achieve the above object, according to one aspect of an embodiment of the present invention, there is provided an information push method, comprising:

[0007] Classify the labels of each transaction subject, so as to build a label system for each transaction subject;

[0008] Obtain the source data corresponding to the label system, extract the high-level features and low-level features of the source data and the label system, and then fuse the high-level features and low-level features to obtain fused features, and build user profiles of each transaction subject based on the fused features;

[0009] The user portraits are matched with the candidate push information for similarity, so as to determine the target push information corresponding to each user portrait from the candidate push information, and the target push information is pushed to the transaction subject corresponding to the corresponding user portrait.

[0010] Optionally, the labels of each transaction subject are classified to construct a label system for each transaction subject, including:

[0011] For each transaction subject, determine the label level of the transaction subject;

[0012] For each label level, determine the next level label corresponding to each label in the label level, so as to build a label system for the transaction subject.

[0013] Optionally, a user profile of each transaction subject is constructed based on the fusion features, including:

[0014] Calculate the feature weights corresponding to the fusion features;

[0015] The fused features are used as nodes, and the feature weights corresponding to the fused features are used as the connections between the nodes. Based on the nodes and connections, the corresponding user profiles of each transaction subject are constructed.

[0016] Optionally, the source data includes at least one of the following:

[0017] Supplier user attribute data, distributor user attribute data, consumer user attribute data, product transaction data, product origin data, product price data, and product testing data.

[0018] Optionally, before constructing the user profile of each transaction subject based on the fusion feature, the method further includes:

[0019] De-duplicate source data, remove invalid data, and remove abnormal data.

[0020] Optionally, a user profile of each transaction subject is constructed based on the fusion features, including:

[0021] Determine the business indicators corresponding to each transaction subject based on the label system;

[0022] The features corresponding to the business indicators are determined from the fused features, and then the user profiles of each transaction subject are constructed based on the features corresponding to the business indicators.

[0023] Optionally, each user profile is matched with the candidate push information by similarity, including:

[0024] Based on the word embedding method, each user portrait is converted into each user portrait vector, and the candidate push information is converted into the candidate information vector;

[0025] The similarity between each user portrait vector and the candidate information vector is calculated, and the candidate push information corresponding to the candidate information vector having a similarity greater than a preset threshold is determined as information matching the user portrait corresponding to the corresponding user portrait vector.

[0026] Optionally, determining the target push information corresponding to each user profile from the candidate push information includes:

[0027] The information matching the user portrait corresponding to the corresponding user portrait vector is determined as the target push information corresponding to the corresponding user portrait.

[0028] In addition, according to another aspect of an embodiment of the present invention, there is provided an information push device, including:

[0029] A label system building unit is configured to classify labels of each transaction subject, thereby building a label system for each transaction subject;

[0030] A user profile building unit is configured to obtain source data corresponding to the label system, extract high-level features and low-level features of the source data and the label system, and then fuse the high-level features and the low-level features to obtain fused features, and build a user profile of each transaction subject based on the fused features;

[0031] The information push unit is configured to perform similarity matching between each user portrait and the candidate push information, so as to determine the target push information corresponding to each user portrait from the candidate push information, and push the target push information to the transaction subject corresponding to the corresponding user portrait.

[0032] Optionally, the label system building unit is further configured to:

[0033] For each transaction subject, determine the label level of the transaction subject;

[0034] For each label level, determine the next level label corresponding to each label in the label level, so as to build a label system for the transaction subject.

[0035] Optionally, the user portrait building unit is further configured to:

[0036] Calculate the feature weights corresponding to the fusion features;

[0037] The fused features are used as nodes, and the feature weights corresponding to the fused features are used as the connections between the nodes. Based on the nodes and connections, the corresponding user profiles of each transaction subject are constructed.

[0038] Optionally, the source data includes at least one of the following:

[0039] Supplier user attribute data, distributor user attribute data, consumer user attribute data, product transaction data, product origin data, product price data, and product testing data.

[0040] Optionally, the information push device further includes a data processing unit configured to:

[0041] De-duplicate source data, remove invalid data, and remove abnormal data.

[0042] Optionally, the user portrait building unit is further configured to:

[0043] Determine the business indicators corresponding to each transaction subject based on the label system;

[0044] The features corresponding to the business indicators are determined from the fused features, and then the user profiles of each transaction subject are constructed based on the features corresponding to the business indicators.

[0045] Optionally, the information push unit is further configured to:

[0046] Based on the word embedding method, each user portrait is converted into each user portrait vector, and the candidate push information is converted into the candidate information vector;

[0047] The similarity between each user portrait vector and the candidate information vector is calculated, and the candidate push information corresponding to the candidate information vector having a similarity greater than a preset threshold is determined as information matching the user portrait corresponding to the corresponding user portrait vector.

[0048] Optionally, the information push unit is further configured to:

[0049] The information matching the user portrait corresponding to the corresponding user portrait vector is determined as the target push information corresponding to the corresponding user portrait.

[0050] According to another aspect of an embodiment of the present invention, there is further provided an electronic device, including:

[0051] one or more processors;

[0052] a storage device for storing one or more programs,

[0053] When one or more programs are executed by one or more processors, the one or more processors implement the information push method of any of the above embodiments.

[0054] According to another aspect of the embodiments of the present invention, a computer-readable medium is provided, on which a computer program is stored. When the program is executed by a processor, the information push method of any of the above embodiments is implemented.

[0055] According to another aspect of an embodiment of the present invention, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the information push method of any of the above embodiments is implemented.

[0056] An embodiment of the above invention has the following advantages or beneficial effects: by classifying the labels of each transaction subject, a label system for each transaction subject is constructed; the source data corresponding to the label system is obtained, the high-level features and low-level features of the source data and the label system are extracted, and then the high-level features and low-level features are fused to obtain fused features, and user portraits of each transaction subject are constructed based on the fused features; each user portrait is matched with the candidate push information for similarity to determine the target push information corresponding to each user portrait from the candidate push information, and the target push information is pushed to the transaction subject corresponding to the corresponding user portrait, so as to provide personalized and targeted information for the different needs and preferences of suppliers, distributors and consumers, thereby reducing the communication costs between suppliers, distributors and consumers, connecting product supply-sales-purchase, and improving product transaction efficiency.

[0057] The further effects of the above-mentioned non-conventional optional manner will be described below in conjunction with specific implementation examples. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0059] Figure 1 is a schematic diagram of the main process of the information push method according to an embodiment of the present invention;

[0060] Figure 2a , 2b and 2c are schematic diagrams of user portraits according to an embodiment of the present invention;

[0061] Figure 3 is a schematic diagram of the main process of the information push method according to an embodiment of the present invention;

[0062] Figure 4 is a schematic diagram of main units of an information push device according to an embodiment of the present invention;

[0063] Figure 5 is an exemplary system architecture diagram to which embodiments of the present invention may be applied;

[0064] Figure 6 It is a schematic diagram of the structure of a computer system of a terminal device or a server suitable for implementing an embodiment of the present invention. DETAILED DESCRIPTION

[0065] The following is a description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, including various details of the embodiments of the present invention to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0066] The acquisition, storage, use, and processing of data in the technical solution of this application comply with the relevant provisions of national laws and regulations.

[0067] Figure 1 is a schematic diagram of the main process of the information push method according to an embodiment of the present application, such as Figure 1 As shown, the information push method includes:

[0068] Step S101, classify the labels of each transaction subject, so as to construct a label system for each transaction subject.

[0069] In this embodiment, the execution subject of the information push method (for example, it can be a server) can obtain the transaction subject with a label through a wired connection or a wireless connection. Among them, the transaction subject includes a supplier, a distributor and a consumer. The labels of each transaction subject are classified, and the classified labels can include basic labels, business ability labels, business content labels, rating labels, preference labels, etc. Among them, the basic label can include information such as name, gender, age, contact information, etc.; the business ability label can include information such as the business subject, business license, legal person, license, business qualification, registered place, business location, etc.; the business content label can include information such as the category of agricultural products, the name of agricultural products, the origin of agricultural products, and the price of agricultural products; the rating label can include information such as business advantages, income contribution, sales volume, and penalties; the preference label can include information such as transaction time preference, purchase location preference, agricultural product origin preference, agricultural product category preference, and agricultural product price preference. The execution subject can also further subdivide the classified labels into primary labels, secondary labels, and tertiary labels, and then generate a label system corresponding to the transaction subject according to the further subdivided primary labels, secondary labels, and tertiary labels. Among them, the second-level label is a refinement of the first-level label, and the third-level label is a refinement of the second-level label.

[0070] After the execution subject classifies the labels of each transaction subject and refines the classified labels to obtain primary labels, secondary labels and tertiary labels, a tree-like label system can be generated based on the primary labels, secondary labels and tertiary labels, wherein the primary labels can be used as the root nodes of the tree-like label system, the secondary labels can be used as the child nodes of the tree-like label system, and the tertiary labels can be used as the leaf nodes of the tree-like label system. The root node is the topmost node of the tree-like label system, with child nodes connected below it; the child node is a node other than the root node, and also has leaf nodes connected below it; the leaf node is a node that is no longer connected to a node below it, that is, an end node.

[0071] Step S102, obtain the source data corresponding to the label system, extract the high-level features and low-level features of the source data and the label system, and then fuse the high-level features and the low-level features to obtain fused features, and build user profiles of various transaction entities based on the fused features.

[0072] Specifically, the source data includes at least one of the following:

[0073] Supplier user attribute data, distributor user attribute data, consumer user attribute data, product transaction data, product origin data, product price data, product testing data. Furthermore, data can be supplemented through industry research, user interviews, user information filling and questionnaires.

[0074] Specifically, before constructing the user profile of each transaction subject based on the fusion features, the method also includes: deduplication of source data, removal of invalid data and removal of abnormal data. In order to ensure data quality, before feature extraction, the source data needs to be deduplicated, invalid data removed and abnormal data removed. Data cleaning is to recheck and verify the data to ensure data consistency.

[0075] In the embodiment of the present application, the extracted high-level features of the source data and the label system can be the semantic information corresponding to the source data and the label system, and specifically can be the corresponding image generated based on the source data and the label system, and the image can be a line chart or a pie chart. Then the execution subject can extract semantic information from the generated image to obtain high-level features. Specifically, the principle of obtaining high-level features is to input the generated image into the feature extraction model, and then call the deep network of the feature extraction model to extract some coarse-grained information, specifically the increase of the receptive field, the increase of the overlapping area between the receptive fields, and the compression of the input image information. What is obtained is some information about the integrity of the image, such as similar features, mutually exclusive features, future trend features, correlation features, etc., so that high-level features can be obtained. The low-level features of the extracted source data and label system can be features such as maximum values ​​and minimum values ​​containing more details. For example, the execution subject can input the generated image into the feature extraction model, and then call the shallow network of the feature extraction model to extract low-level features containing more pixel information (such as maximum values ​​and minimum values), and the overlapping area of ​​the receptive field is small, so as to ensure that the shallow network captures more details.

[0076] After extracting high-level features and low-level features, the execution subject can fuse the high-level features and low-level features based on the canonical correlation analysis (CCA) method. The fusion method based on CCA (canonical correlation analysis) uses the correlation between two input features (high-level features and low-level features). The features obtained after two transformations have a higher correlation than the two input features. The purpose of feature fusion is to merge the high-level features and low-level features extracted from the image into a more discriminative feature.

[0077] Specifically, a user profile of each transaction subject is constructed based on the fused features, including: calculating the feature weights corresponding to the fused features. Specifically, the execution subject can call the TF / IDF algorithm to calculate the feature weights corresponding to the fused features; taking the fused features as nodes, and the feature weights corresponding to the fused features as lines between the nodes, and constructing the corresponding user profiles of each transaction subject based on the nodes and lines.

[0078] As another implementation method of the embodiment of the present application, the execution subject can also construct the user portrait of each transaction subject in the following manner: for each transaction subject, the corresponding source data is obtained; according to the source data and label system corresponding to each transaction subject, the source data and label system are extracted to construct the user portrait of each transaction subject. For each transaction subject, the source data of the user is first obtained from the database, and then the transaction subject to which each source data belongs is determined, such as the supplier, distributor or consumer, and then the source data of the transaction subject is extracted to construct the user portrait of the transaction subject according to the label system of the transaction subject.

[0079] For example, if the transaction subject is a supplier, such as an agricultural product supplier, feature extraction is performed on the source data of the transaction subject according to the label system of the agricultural product supplier, thereby constructing a user profile of the agricultural product supplier, such as Figure 2a Similarly, if the transaction subject is a distributor, such as an agricultural product distributor, the constructed agricultural product distributor user profile is as follows: Figure 2b As shown; if the transaction subject is a consumer, such as an agricultural product consumer, the constructed agricultural product consumer user portrait is as follows Figure 2c shown.

[0080] The embodiment of the present application constructs user portraits for three different transaction entities, namely, suppliers, distributors, and consumers. In fact, it is to label the user information of different transaction entities, that is, to abstract a labeled user model based on the basic attributes, business capabilities, ratings, preferences and other information of different users, so as to accurately push information to the corresponding transaction entities based on the constructed user portraits of each transaction entity, thereby reducing the communication costs between the various transaction entities, connecting product supply-sales-purchase, and improving product transaction efficiency.

[0081] Specifically, a user profile of each transaction subject is constructed based on the fused features, including: determining the business indicators corresponding to each transaction subject according to the label system; determining the features corresponding to the business indicators from the fused features, and then constructing a user profile of each transaction subject based on the features corresponding to the business indicators.

[0082] For example, business indicators may include price, origin, distance between the sales location and the delivery location, quality, water content, etc. The embodiments of the present application do not specifically limit the content of business indicators. The execution entity may filter out similar features from the fusion features based on the business indicators to support the construction of user portraits of various transaction entities. The specific screening method may be: converting the business indicator and each fusion feature into a vector, calculating the similarity between the vector corresponding to the business indicator and the vector corresponding to each fusion feature, determining the fusion feature corresponding to the vector corresponding to the fusion feature whose similarity to the vector corresponding to the business indicator is greater than a preset similarity threshold as the target fusion feature, and then constructing the corresponding user portrait of the transaction entity based on the target fusion feature.

[0083] Step S103, performing similarity matching between each user portrait and the candidate push information, so as to determine the target push information corresponding to each user portrait from the candidate push information, and push the target push information to the transaction subject corresponding to the corresponding user portrait.

[0084] Specifically, each user portrait is matched with the candidate push information by similarity, including: converting each user portrait into a user portrait vector and converting the candidate push information into a candidate information vector based on a word embedding method; calculating the similarity between each user portrait vector and the candidate information vector, and determining the candidate push information corresponding to the candidate information vector having a similarity greater than a preset threshold as the information that matches the user portrait corresponding to the corresponding user portrait vector.

[0085] Specifically, determining the target push information corresponding to each user portrait from the candidate push information includes: determining the information matching the user portrait corresponding to the corresponding user portrait vector as the target push information corresponding to the corresponding user portrait.

[0086] Specifically, the candidate push information may include supplier information, store information, product discount information, product quality information, product origin information, etc. The embodiment of the present application does not specifically limit the content of the candidate push information.

[0087] It is understandable that the candidate push information can be suitable for pushing to each transaction subject. The execution subject can filter out the information corresponding to each transaction subject from the candidate push information based on the user profile of each transaction subject, and push the filtered information as the target push information to the corresponding transaction subject. In this way, personalized information push can be carried out to each transaction subject, improving product transaction efficiency.

[0088] This embodiment constructs a label system for each transaction entity by classifying the labels of each transaction entity; obtains source data corresponding to the label system, extracts high-level features and low-level features of the source data and the label system, and then fuses the high-level features and low-level features to obtain fused features, and constructs user portraits of each transaction entity based on the fused features; performs similarity matching between each user portrait and candidate push information to determine the target push information corresponding to each user portrait from the candidate push information, and pushes the target push information to the transaction entity corresponding to the corresponding user portrait, so as to provide personalized and targeted information for the different needs and preferences of suppliers, distributors, and consumers, thereby reducing the communication costs among suppliers, distributors, and consumers, connecting product supply-sales-purchase, and improving product transaction efficiency.

[0089] Figure 3 It is a schematic diagram of the main process of the information push method according to an embodiment of the present invention.

[0090] like Figure 3 As shown, the information push method includes:

[0091] Step S301: for each transaction subject, determine the label level of the transaction subject.

[0092] Specifically, the tag level may include a primary tag, a secondary tag, and a tertiary tag.

[0093] Step S302: for each label level, determine the next level label corresponding to each label in the label level, so as to construct a label system for the transaction subject.

[0094] For example, for each label level, the second-level label is the label of the next level of the first-level label, and the third-level label is the label of the next level of the second-level label. The second-level label is a refinement of the first-level label, and the third-level label is a refinement of the second-level label. The constructed label system of the transaction subject can also be displayed in a table.

[0095] For example, taking the transaction subject as the supplier, the constructed label system is shown in Table 1 below:

[0096] Table 1 Labeling system of suppliers

[0097]

[0098]

[0099] For example, taking the transaction subject as the distributor, the constructed label system is shown in Table 2 below:

[0100] Table 2 Labeling system of distributors

[0101]

[0102]

[0103]

[0104] Similarly, a label system for consumers is constructed, as shown in Table 3 below:

[0105] Table 3 Labeling system for consumers

[0106]

[0107]

[0108] In the scenario of "supply-sale-purchase" of agricultural products, the "supply-sale-purchase" chain is a comprehensive ecology. The main subjects of product transaction user portraits include suppliers, distributors, and consumers. Profiling different transaction subjects is conducive to connecting the agricultural "supply-sale-purchase" chain in series and reducing the communication costs between different transaction subjects. Therefore, in this step, the embodiment of the present invention first classifies the labels of each transaction subject, and then constructs the label system of each transaction subject according to the label classification results of each transaction subject.

[0109] Step S303, obtain the source data corresponding to the label system, extract the high-level features and low-level features of the source data and the label system, and then fuse the high-level features and the low-level features to obtain fused features, and build user profiles of various transaction entities based on the fused features.

[0110] The source data corresponding to the tag system may specifically be the source data corresponding to each tag in the tag system. The source data may be, for example, the name, gender, age, contact information, etc. corresponding to the basic tag.

[0111] By constructing a user profile through feature fusion, the constructed user profile can better reflect the demand characteristics of the corresponding transaction subject, so that the required information can be pushed to the corresponding transaction subject more accurately based on the constructed user profile.

[0112] Step S304, performing similarity matching between each user portrait and the candidate push information, so as to determine the target push information corresponding to each user portrait from the candidate push information, and push the target push information to the transaction subject corresponding to the corresponding user portrait.

[0113] The embodiments of the present application can provide personalized and targeted information according to the different needs and preferences of suppliers, distributors, and consumers, thereby reducing the communication costs between suppliers, distributors, and consumers, connecting product supply-sales-purchase, and improving product transaction efficiency.

[0114] As an example of the present application, in the agricultural product supply and marketing scenario, by positioning the subject of the transaction (including agricultural product suppliers, agricultural product distributors, and agricultural product consumers) in the "supply-sales-purchase" chain of the agricultural wholesale market, the labels are clustered according to the label characteristics of different transaction subjects, and the different categories of information clustered according to the subject characteristics are subdivided to form specific information labels, and a composite algorithm model is used for accurate matching. Through intelligent recommendation, the different needs and preferences of agricultural product suppliers, agricultural product distributors, and agricultural product consumers are achieved, and personalized and targeted agricultural trade information is provided, thereby reducing the communication costs between suppliers, distributors, and consumers, connecting the supply-sales-purchase of agricultural products in series, and improving the transaction efficiency of agricultural products in the agricultural wholesale market. In the embodiment of the present application, the agricultural product supplier is the supplier, the agricultural product distributor is the distributor, and the agricultural product consumer is the consumer.

[0115] In the agricultural and industrial product supply and marketing scenario, the information push method may include the following steps:

[0116] The first step: locate the subject of the portrait. The "supply-sales-purchase" chain of the agricultural wholesale market is a very comprehensive ecosystem. The subject of the portrait of agricultural product trading users includes agricultural product suppliers, agricultural product distributors, and agricultural product consumers. Profiling different subjects is conducive to connecting the "supply-sales-purchase" chain of the agricultural wholesale market and reducing the communication costs between different trading subjects.

[0117] Step 2: Information labeling. Collecting source data from the smart agricultural wholesale platform of the agricultural wholesale market requires considering multiple dimensions of data, including: market agricultural product supplier user attribute data, market agricultural product dealer user attribute data, market agricultural product consumer user attribute data, market farm product transaction data, market agricultural product origin data, market agricultural product price data, market agricultural product testing data, and supplementing data through industry surveys, user interviews, user information filling and questionnaires. Cluster the labels according to the label characteristics of different subjects, and form specific information labels after subdividing the different categories of information clustered according to the subject characteristics. For example, basic labels include name, gender, ID number, contact information, etc.; business ability labels include business entity, business license, legal person, license, business qualification, registered place, business premises, etc.; business content labels include business agricultural product categories, agricultural product names, agricultural product origins, agricultural product prices, etc.; rating labels include business advantages, income contribution, sales volume, penalties, etc.; preference labels include transaction time preferences, purchase location preferences, agricultural product origin preferences, agricultural product category preferences, agricultural product price preferences, etc.

[0118] Step 3: Build user portraits for different transaction entities. Profiling the three different transaction entities, namely, agricultural product suppliers, agricultural product distributors, and agricultural product consumers, is actually labeling the user information of different transaction entities, that is, a labeled user model is abstracted based on the basic attributes, business capabilities, ratings, preferences, and other information of different users.

[0119] Step 4: Intelligent recommendation of agricultural trade information. Push agricultural trade information to each trading entity based on the user portrait of each trading entity and the attribute information of each product. Push agricultural trade information (such as product name, product price, product origin, high-quality distributors, high-quality supplier information, etc.) to each trading entity based on the user portrait of each trading entity and the attribute information of each product (such as price, origin, category, etc.) that have been constructed. Specifically: Based on the user portrait of each trading entity and the attribute information of each product, and using a composite recommendation algorithm, calculate the agricultural trade information corresponding to each trading entity respectively; push agricultural trade information to each trading entity respectively; wherein the agricultural trade information includes at least one of the following: product information, user information corresponding to the supplier, and user information corresponding to the distributor. After profiling each trading entity corresponding to the supplier, distributor, and consumer and labeling the information, a composite algorithm model is used for precise matching, so as to provide personalized and targeted agricultural trade information for the different needs and preferences of different trading entities. For example: when the transaction subject is the supplier, the company provides the supplier with information on high-quality distributors and products of origin, etc.; when the transaction subject is the distributor, the company provides the distributor with precision marketing, store marketing, high-quality supplier information, etc.; when the transaction subject is the consumer, the company provides the consumer with agricultural wholesale discount information, real-time product price information, high-quality product information, etc.

[0120] Optionally, the composite recommendation algorithm includes at least two of the following: a recommendation algorithm based on association rules, a collaborative filtering recommendation algorithm, and a content-based recommendation algorithm. The recommendation algorithm based on association rules seeks the correlation between data sets based on data analysis. First, it finds the correlation between frequently occurring items in the event, forms an association rule in the form of X→Y, obtains the event of interest to the transaction subject, and then makes intelligent recommendations. The association recommendation algorithm belongs to an unsupervised learning method. The collaborative filtering recommendation algorithm studies the historical behavior of the transaction subject, finds transaction subjects with similar behaviors or events with common interests, finds the correlation between transaction subjects or events, and then implements recommendations based on the correlation. The collaborative filtering algorithm is mainly divided into a collaborative filtering algorithm based on the transaction subject, a collaborative filtering algorithm based on the project, and a collaborative filtering algorithm based on the model. The content-based recommendation algorithm calculates the similarity between the user portrait of each transaction subject and the candidate push information, thereby implementing the recommendation. The machine learning methods used in the recommendation process include decision trees, neural networks, etc.

[0121] According to the various embodiments of the present application, it can be seen that the embodiments of the present invention first classify the labels of trading entities such as suppliers, distributors and consumers, thereby constructing a label system for each trading entity, and then constructing a user portrait of each trading entity based on the label system of each trading entity and the source data of each trading entity. Finally, the technical means of pushing agricultural trade information to each trading entity based on the user portrait of each trading entity and the attribute information of each product solves the technical problem that the information communication between suppliers, distributors and consumers in the prior art is not smooth and accurate agricultural trade information cannot be obtained. The embodiments of the present invention construct a label system for trading entities in the "supply-sale-purchase" chain of agricultural wholesale, and then extract features from the source data of each trading entity, thereby constructing a user portrait of each trading entity, so as to push agricultural trade information to each trading entity, realize the different needs and different preferences of each trading entity (such as suppliers, distributors, and consumers), and provide personalized and targeted agricultural trade information, thereby reducing the communication costs between suppliers, distributors, and consumers, connecting product supply-sale-purchase in series, and improving the transaction efficiency of agricultural wholesale products.

[0122] Figure 4 Schematic diagram of the main units of the information push device according to the embodiment of the present application. Figure 4 As shown, the information pushing device 400 includes a label system building unit 401, a user portrait building unit 402 and an information pushing unit 403.

[0123] The label system construction unit 401 is configured to classify labels of various transaction entities, thereby constructing a label system for each transaction entity.

[0124] The user profile construction unit 402 is configured to obtain source data corresponding to the label system, extract high-level features and low-level features of the source data and the label system, and then perform feature fusion on the high-level features and the low-level features to obtain fused features, and construct user profiles of each transaction subject based on the fused features.

[0125] The information push unit 403 is configured to perform similarity matching between each user portrait and the candidate push information, so as to determine the target push information corresponding to each user portrait from the candidate push information, and push the target push information to the transaction subject corresponding to the corresponding user portrait.

[0126] In some embodiments, the label system construction unit 401 is further configured to: for each transaction subject, determine the label level of the transaction subject; for each label level, determine the next level label corresponding to each label in the label level, thereby constructing the label system of the transaction subject.

[0127] In some embodiments, the user portrait construction unit 402 is further configured to: calculate the feature weights corresponding to the fused features; use the fused features as nodes, and the feature weights corresponding to the fused features as lines between the nodes, and construct corresponding user portraits of each transaction subject based on the nodes and lines.

[0128] In some embodiments, the source data includes at least one of the following: supplier user attribute data, distributor user attribute data, consumer user attribute data, product transaction data, product origin data, product price data, and product testing data.

[0129] In some embodiments, the information push device further includes Figure 4 The data processing unit not shown in the figure is configured to: deduplicate the source data, remove invalid data and remove abnormal data.

[0130] In some embodiments, the user portrait construction unit 402 is further configured to: determine the business indicators corresponding to each transaction entity based on the label system; determine the features corresponding to the business indicators from the fused features, and then construct the user portrait of each transaction entity based on the features corresponding to the business indicators.

[0131] In some embodiments, the information push unit 403 is further configured to: convert each user portrait into each user portrait vector and convert the candidate push information into a candidate information vector based on a word embedding method; calculate the similarity between each user portrait vector and the candidate information vector, and determine the candidate push information corresponding to the candidate information vector whose similarity is greater than a preset threshold as information that matches the user portrait corresponding to the corresponding user portrait vector.

[0132] In some embodiments, the information push unit 403 is further configured to: determine the information matching the user portrait corresponding to the corresponding user portrait vector as the target push information corresponding to the corresponding user portrait.

[0133] It should be noted that the information push method and the information push device of the present application have a corresponding relationship in terms of specific implementation content, so the repeated content will not be described again.

[0134] Figure 5 An exemplary system architecture 500 is shown to which the information push method or information push device according to the embodiments of the present application can be applied.

[0135] like Figure 5As shown, system architecture 500 may include terminal devices 501, 502, 503, a network 504 and a server 505. Network 504 is used to provide a medium for communication links between terminal devices 501, 502, 503 and server 505. Network 504 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0136] Users can use terminal devices 501, 502, 503 to interact with server 505 through network 504 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 501, 502, 503, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (only as examples).

[0137] The terminal devices 501, 502, 503 may be various electronic devices having an information push processing screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, and the like.

[0138] Server 505 may be a server that provides various services, such as a background management server that provides processing support for each transaction subject to be labeled obtained by the user using terminal devices 501, 502, and 503 (only as an example). The background management server can classify the labels of each transaction subject to build a label system for each transaction subject; obtain the source data corresponding to the label system, extract the high-level features and low-level features of the source data and the label system, and then fuse the high-level features and low-level features to obtain fusion features, and build user portraits of each transaction subject based on the fusion features; match each user portrait with the candidate push information for similarity, so as to determine the target push information corresponding to each user portrait from the candidate push information, and push the target push information to the transaction subject corresponding to the corresponding user portrait, so as to provide personalized and targeted information for the different needs and preferences of suppliers, distributors, and consumers, thereby reducing the communication costs between suppliers, distributors, and consumers, connecting product supply-sales-purchase in series, and improving product transaction efficiency.

[0139] It should be noted that the information push method provided in the embodiment of the present application is generally executed by the server 505 , and accordingly, the information push device is generally arranged in the server 505 .

[0140] It should be understood that Figure 5 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0141] Reference below Figure 6, which shows a schematic diagram of the structure of a computer system 600 of a terminal device suitable for implementing an embodiment of the present application. Figure 6 The terminal device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0142] like Figure 6 As shown, the computer system 600 includes a central processing unit (CPU) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage part 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the computer system 600 are also stored. The CPU 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0143] The following components are connected to the I / O interface 605: an input section 606 including a keyboard, a mouse, etc.; an output section 607 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN card, a modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the I / O interface 605 as needed. A removable medium 611, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 610 as needed, so that a computer program read therefrom is installed into the storage section 608 as needed.

[0144] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 609, and / or installed from the removable medium 611. When the computer program is executed by the central processing unit (CPU) 601, the above-mentioned functions defined in the system of the present application are executed.

[0145] It should be noted that the computer-readable medium shown in the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. Computer-readable storage media may include, for example, but are not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0146] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the above-mentioned module, program segment or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0147] The units involved in the embodiments described in the present application may be implemented by software or by hardware. The units described may also be set in a processor, for example, it may be described as: a processor includes a label system construction unit, a user portrait construction unit, and an information push unit. Among them, the names of these units do not constitute a limitation on the units themselves in some cases.

[0148] As another aspect, the present application also provides a computer-readable medium, which may be included in the device described in the above embodiment; or it may exist independently and not be assembled into the device. The above computer-readable medium carries one or more programs. When the above one or more programs are executed by a device, the device classifies the labels of each transaction subject, thereby building a label system for each transaction subject; obtains the source data corresponding to the label system, extracts the high-level features and low-level features of the source data and the label system, and then fuses the high-level features and low-level features to obtain fusion features, and builds user portraits of each transaction subject based on the fusion features; matches each user portrait with the candidate push information for similarity, so as to determine the target push information corresponding to each user portrait from the candidate push information, and pushes the target push information to the transaction subject corresponding to the corresponding user portrait, so as to provide personalized and targeted information for the different needs and preferences of suppliers, distributors, and consumers, thereby reducing the communication costs between suppliers, distributors, and consumers, connecting product supply-sales-purchase in series, and improving product transaction efficiency.

[0149] The computer program product of the present application includes a computer program, and when the computer program is executed by a processor, the information push method in the embodiment of the present application is implemented.

[0150] According to the technical solution of the embodiment of the present application, it is possible to provide personalized and targeted information according to the different needs and preferences of suppliers, distributors and consumers, thereby reducing the communication costs between suppliers, distributors and consumers, connecting product supply-sales-purchase, and improving product transaction efficiency.

[0151] The above specific implementations do not constitute a limitation on the protection scope of this application. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may occur depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of this application should be included in the protection scope of this application.

Claims

1. An information push method, characterized in that: include: Classify the labels of each transaction subject, so as to construct a label system for each transaction subject; Acquire source data corresponding to the label system, extract high-level features and low-level features of the source data and the label system, and then perform feature fusion on the high-level features and the low-level features to obtain fused features; Constructing user profiles of the respective transaction entities based on the fused features, including: determining business indicators corresponding to the respective transaction entities according to the tag system; determining features corresponding to the business indicators from the fused features, and then constructing user profiles of the respective transaction entities based on the features corresponding to the business indicators; The user portraits are matched with the candidate push information for similarity, so as to determine the target push information corresponding to each user portrait from the candidate push information, and the target push information is pushed to the transaction subject corresponding to the corresponding user portrait.

2. The method according to claim 1, characterized in that: Classify the labels of each transaction subject, so as to construct a label system for each transaction subject, including: For each transaction subject, determining a label level of the transaction subject; For each label level, the labels of the next level corresponding to each label in the label level are determined, so as to construct a label system for the transaction subject.

3. The method according to claim 1, characterized in that The constructing the user profile of each transaction subject based on the fusion feature includes: Calculating feature weights corresponding to the fused features; The fused features are used as nodes, and the feature weights corresponding to the fused features are used as lines between the nodes, and user profiles of the corresponding transaction entities are constructed based on the nodes and the lines.

4. The method according to claim 1, characterized in that: The source data includes at least one of the following: Supplier user attribute data, distributor user attribute data, consumer user attribute data, product transaction data, product origin data, product price data, and product testing data.

5. The method according to claim 1, characterized in that: Before constructing the user portraits of the respective transaction entities based on the fusion features, the method further includes: The source data is deduplicated, invalid data is removed, and abnormal data is removed.

6. The method according to claim 1, characterized in that The similarity matching of each user portrait with the candidate push information includes: Based on the word embedding method, each of the user portraits is converted into a user portrait vector, and the candidate push information is converted into a candidate information vector; The similarity between each user portrait vector and the candidate information vector is calculated, and the candidate push information corresponding to the candidate information vector having a similarity greater than a preset threshold is determined as information matching the user portrait corresponding to the corresponding user portrait vector.

7. The method according to claim 6, characterized in that The determining the target push information corresponding to each of the user portraits from the candidate push information includes: The information matching the user portrait corresponding to the corresponding user portrait vector is determined as the target push information corresponding to the corresponding user portrait.

8. An information push device, characterized in that: include: A label system construction unit is configured to classify labels of various transaction entities, thereby constructing a label system for each transaction entity; The user profile construction unit is configured to obtain source data corresponding to the label system, extract high-level features and low-level features of the source data and the label system, and then perform feature fusion on the high-level features and low-level features to obtain fusion features, and construct user profiles of each transaction subject based on the fusion features, including: determining business indicators corresponding to each transaction subject according to the label system; determining features corresponding to the business indicators from the fusion features, and then constructing user profiles of each transaction subject based on the features corresponding to the business indicators; The information push unit is configured to perform similarity matching between each of the user portraits and the candidate push information, so as to determine the target push information corresponding to each of the user portraits from the candidate push information, and push the target push information to the transaction subject corresponding to the corresponding user portrait.

9. The device according to claim 8, characterized in that The label system building unit is further configured to: For each transaction subject, determining a label level of the transaction subject; For each label level, the labels of the next level corresponding to each label in the label level are determined, so as to construct a label system for the transaction subject.

10. The device according to claim 8, characterized in that The user portrait building unit is further configured to: Calculating feature weights corresponding to the fused features; The fused features are used as nodes, and the feature weights corresponding to the fused features are used as lines between the nodes, and user profiles of the corresponding transaction entities are constructed based on the nodes and the lines.

11. The device according to claim 8, characterized in that The source data includes at least one of the following: Supplier user attribute data, distributor user attribute data, consumer user attribute data, product transaction data, product origin data, product price data, and product testing data.

12. The device according to claim 8, characterized in that The device further comprises a data processing unit configured to: The source data is deduplicated, invalid data is removed, and abnormal data is removed.

13. An electronic device, characterized in that: include: one or more processors; a storage device 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 method according to any one of claims 1 to 7.

14. A computer readable medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

15. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images