Agricultural supply and demand information matching method and system

Abstract

The present invention relates to the technical field of agricultural supply and demand information matching, and provides a method and system for agricultural supply and demand information matching. The agricultural supply and demand information matching method includes constructing a knowledge graph in a database with products as core atoms, assigning attributes to the core atoms, and associating all agricultural supply and demand information associated with the core atoms in the knowledge graph; determining role attributes and extracting the subject and entity of the input search text respectively; establishing matching rules, and using different matching rules to search and match in the database based on the role attributes and whether the subject and entity in the search text exist to obtain matching results. The agricultural supply and demand information matching method and system provided by the present invention can provide accurate and convenient search and matching services for agricultural supply and demand information demanders, solve the problem of manual review of matching results in the existing technology, improve the accuracy of matching results during search and matching, and enable the effective utilization of massive amounts of agricultural supply and demand information.

Description

Technical Field

[0001] The present invention relates to the technical field of agricultural supply and demand information matching, and in particular to an agricultural supply and demand information matching method and system. Background Art

[0002] In the current agricultural Internet, agricultural supply and demand information matching is still at a relatively backward stage. In order to facilitate the acquisition of the required agricultural supply and demand information, the existing agricultural supply and demand information matching method is keyword retrieval matching, that is, searching by entering keywords into the retrieval system, and the retrieval system matches the agricultural supply and demand information associated with the keywords in the database based on the input keywords.

[0003] However, the keyword search and matching method often requires using multiple keywords to search and collect massive amounts of agricultural supply and demand information in the database, which will lead to inaccurate matching results. At the same time, the accuracy of the matching results needs to be manually reviewed again, resulting in low search and matching efficiency. In addition, most of the agricultural supply and demand information is sunk in the database, and good resources cannot be used in a timely manner. Summary of the Invention

[0004] The purpose of the present invention is to provide an agricultural supply and demand information matching method to at least overcome the technical problems in the existing technology caused by keyword retrieval and matching, such as inaccurate matching results, low retrieval and matching efficiency, the need for manual review of matching results, and the inability of most agricultural supply and demand information to be used in a timely manner.

[0005] The purpose of the present invention is achieved through the following technical solutions:

[0006] In one aspect, the present invention provides a method for matching agricultural supply and demand information, comprising the following steps:

[0007] Step S1. Construct a knowledge graph of agricultural supply and demand information in the database with products as core atoms, and assign attributes to the core atoms. The knowledge graph associates all agricultural supply and demand information related to the core atoms;

[0008] Step S2. Determine the role attributes and extract the subject and entity of the input search text respectively;

[0009] Step S3. Establish matching rules, and use different matching rules to search for matches in the database according to the role attributes and whether the subject and entity in the search text exist to obtain matching results.

[0010] In some possible embodiments, in step S3, the matching rule includes:

[0011] If there is no subject but an entity, the entity of the extracted search text is placed in the knowledge graph for positioning, and the entity level is determined, whether it is a core atom, an entity in the core atom attribute, or both. Then, based on different judgment results, a match is searched in the database;

[0012] There are topics and entities. The topic of the extracted search text is used as the basis. In the subsequent search and matching process, the agricultural supply and demand information that is most closely related to the topic attributes of the search text is prioritized. Secondly, the entities of the extracted search text are placed in the knowledge graph for positioning. The entity's level is determined, whether it is a core atom or an entity within the core atom's attributes. Then, based on different judgment results, the database is searched for matches.

[0013] There is no subject and no entity. Keyword search and matching is used to search and match the search text with the agricultural supply and demand information in the database. The detailed content and summary of the searched agricultural supply and demand information are then calculated to perform a similarity calculation with the search text for a second search and match. The matching results are sorted by similarity to filter out matching results with a similarity of less than 60%.

[0014] In some possible embodiments, for a case where there is no subject but an entity, searching for a match in the database according to different judgment results is specifically as follows:

[0015] If only the core atom is located, the matching agricultural supply and demand information will be retrieved in the database according to the attributes of the core atom; if only the entity in the core atom attribute is located, the matching core atom associated with the entity in the core atom attribute will be retrieved, and then the matching agricultural supply and demand information will be retrieved in the database in combination with the core atom attribute and the core atom, and the matching results will be sorted according to the release time and validity period of the agricultural supply and demand information; if the core atom and the entity in the core atom attribute are located at the same time, the position of all entities will be located in the knowledge graph first, and the core atom attributes therein will be inferred through the knowledge graph, and then the entities of the retrieved text will be reorganized into new text according to the core atom attributes, and the similarity between the new text and the relevant agricultural supply and demand information in the database will be calculated and sorted.

[0016] In some possible embodiments, for the case where there are both subjects and entities, searching for matches in the database according to different judgment results is specifically as follows:

[0017] If only the core atom is located, the subject of the search text is used as the inference condition, and the relevant agricultural supply and demand information is retrieved and matched in the database. The similarity between the retrieved and matched agricultural supply and demand information and the search text is calculated and sorted, and the validity period of the agricultural supply and demand information is added to the sorting calculation process when sorting the agricultural supply and demand information; if only the entity in the core atom attribute is located, the attribute of the entity is retrieved and matched with the subject of the search text. If the match is successful, it is associated with the core atom through the knowledge graph, and the associated core atom replaces the corresponding entity in the search text. Then the replaced new text is calculated and sorted with the retrieved and matched agricultural supply and demand information.

[0018] In some possible embodiments, in step S1 , the attributes assigned to the core atom are divided into technical attributes, consumption attributes, equipment attributes, and information attributes.

[0019] In some possible embodiments, in step S2, the role attributes are divided into supply and demand information publisher and supply and demand information demander;

[0020] Supply and demand information publishers can enter agricultural supply and demand information into the database. When entering agricultural supply and demand information, they can match the relevant information in the database according to the detailed content of the agricultural supply and demand information, and can regularly update and match the new agricultural supply and demand information released by other supply and demand information publishers;

[0021] Supply and demand information demanders can enter search text to search for matching relevant agricultural supply and demand information in the database.

[0022] In some possible embodiments, in step S2, the topics of the retrieved text are extracted using a text topic classification technique, and the entities of the retrieved text are extracted using an entity recognition technique.

[0023] On the other hand, the present invention provides an agricultural supply and demand information matching system for implementing the agricultural supply and demand information matching method described above. The agricultural supply and demand information matching system is divided into functional modules, including:

[0024] The knowledge graph management module is used to construct a knowledge graph of agricultural supply and demand information in the database with products as core atoms, and assign attributes to the core atoms. The knowledge graph will associate all agricultural supply and demand information related to the core atoms;

[0025] The retrieval and matching module is used to determine role attributes and establish matching rules. The retrieval and matching module retrieves and matches relevant agricultural supply and demand information in the database using different matching rules based on the input search text and displays the matching results.

[0026] In some possible embodiments, the agricultural supply and demand information matching system further includes:

[0027] The supply and demand information entry module is used to enter agricultural supply and demand information into the database, and at the same time structure the data of agricultural supply and demand information and assign agricultural supply and demand information labels.

[0028] The technical solutions of the embodiments of the present invention have at least the following advantages and beneficial effects:

[0029] The agricultural supply and demand information matching method and system provided by the present invention can provide accurate and convenient retrieval and matching services for those who need agricultural supply and demand information, solve the problem of manual review of matching results in the existing technology, improve the accuracy of matching results during retrieval and matching, and enable massive agricultural supply and demand information to be effectively utilized. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 A flowchart of the agricultural supply and demand information matching method provided by an embodiment of the present invention;

[0031] Figure 2 This is a structural block diagram of the agricultural supply and demand information matching system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0032] After analyzing the massive amount of agricultural supply and demand information in existing databases, it was found that the existing agricultural supply and demand information is characterized by a large scope, rich content and strong correlation, and the vast majority of agricultural supply and demand information is stored in the database in the form of text. To this end, this embodiment provides an agricultural supply and demand information matching method to at least overcome the technical problems of inaccurate matching results, low retrieval and matching efficiency, the need for manual review of matching results, and the inability of most agricultural supply and demand information to be used in a timely manner brought about by the use of keyword retrieval and matching in the existing technology. This makes it convenient for agricultural supply and demand information demanders to analyze and retrieve matching from multiple angles when they want to obtain corresponding agricultural supply and demand information, thereby providing agricultural supply and demand information demanders with a convenient and accurate agricultural supply and demand information retrieval and matching process. For example, when agricultural supply and demand information demanders want to retrieve and match agricultural supply and demand information about "Red Fuji Apple", the retrieved and matched agricultural supply and demand information may be seed information related to "Red Fuji Apple", or it may be a studio that provides "Red Fuji Apple" pest and disease control technology, or it may be pesticides and fertilizers needed for growing "Red Fuji Apple", or it may be equipment and instruments needed in the "Red Fuji Apple" planting process.

[0033] For details, please refer to Figure 1 , the agricultural supply and demand information matching method includes the following steps:

[0034] Step S1. Construct a knowledge graph of agricultural supply and demand information in the database with products as core atoms, and assign attributes to the core atoms. The knowledge graph will associate all agricultural supply and demand information related to the core atoms.

[0035] It can be understood that the product mentioned in this embodiment refers to a specific product in the agricultural field, such as "Red Fuji Apple", "Huanghuali", "Emperor Banana", etc. The product is used as the core atom because the types of products are limited and exhaustible. This not only conforms to the usage habits of those who need agricultural supply and demand information, but also shortens the time consumed by retrieval and matching, and improves the accuracy of retrieval and matching.

[0036] At the same time, in step S1, the attributes assigned to the core atom are divided into technical attributes, consumable attributes, equipment attributes and information attributes, among which technical attributes refer to technologies related to pest and disease control, product post-processing and manufacturing, and breeding and cultivation services; consumable attributes refer to consumables such as pesticides and fertilizers required in the planting and cultivation process; equipment attributes refer to agricultural machinery, instruments and other tools and equipment used in the planting and cultivation process; information attributes refer to the basic information of the product, including product alias, department and place of origin.

[0037] At this time, all agricultural supply and demand information associated with the core atom refers to agricultural supply and demand information that meets a certain attribute of the core atom. For example, taking the product "Red Fuji Apple" as the core atom, if the agricultural supply and demand information "Apple Pest and Disease Control Technology" exists in the database, then the agricultural supply and demand information is considered to meet the technical attributes of the core atom and is associated with the core atom; for example, by searching for phosphate fertilizer in pesticides and fertilizers in the knowledge graph, you can find core atoms such as "jasmine" and "gardenia" whose consumption attributes include phosphate fertilizer.

[0038] Step S2: When a search text is input for search matching, role attributes are determined and the subject and entity of the input search text are extracted respectively.

[0039] Specifically, in step S2, the role attributes are divided into supply and demand information publishers and supply and demand information demanders. Supply and demand information publishers can enter agricultural supply and demand information into the database. When entering agricultural supply and demand information, they match the relevant information in the database based on the detailed content of the agricultural supply and demand information. They can also regularly update and match new agricultural supply and demand information released by other supply and demand information publishers to add new agricultural supply and demand information to the database or update the agricultural supply and demand information in the database, thereby better meeting the needs of agricultural supply and demand information demanders. Supply and demand information demanders can enter search text to search the database for agricultural supply and demand information that meets their needs.

[0040] At the same time, the subject of the input search text is extracted using the text subject classification technology, and the entity of the input search text is extracted using the entity recognition technology. It should be noted that the subject classification technology and entity recognition technology used in this embodiment are both conventional technologies in natural language processing technology in the field of artificial intelligence. Among them, the text subject classification technology is to analyze the text as a whole, and at the same time calculate and assign the text one or more existing topics in the subject library; the entity recognition technology is to extract all entities in the text. Entities usually refer to things that exist objectively and can be distinguished from each other, and there may be inclusion relationships between different entities. That is to say, entities are usually nouns in a text. In this embodiment, entities refer to the names of a series of products such as "Red Fuji Apple", "Huanghuali" or "Emperor Banana", or the names of a series of tools and equipment such as agricultural machinery and instruments, etc.

[0041] Step S3. Establish matching rules, and use different matching rules to search for matches in the database according to the role attributes and whether the subject and entity in the search text exist to obtain matching results.

[0042] It is understandable that, under normal circumstances, retrieval and matching will only be performed by inputting search text when the role attribute is an agricultural supply and demand information demander. For example, when the search text "How to plant Red Fuji apples" is input for retrieval and matching, based on the retrieval and matching being performed in an input manner, it can be confirmed that the role attribute this time is an agricultural supply and demand information demander. Secondly, through the subject classification technology, it can be obtained that the subject of the retrieval text contains the technical attributes of the core atoms (i.e., "how to plant"), and then the entity "Red Fuji apple" is extracted through the entity recognition technology. Finally, the extracted entity is searched and matched with the relevant agricultural supply and demand information in the database according to the corresponding matching rules.

[0043] In step S3, the matching rules include three cases: no subject but entity, subject and entity, and no subject and no entity.

[0044] For the first case, when there is no subject but an entity, during search and matching, the entity of the extracted search text is placed in the knowledge graph for positioning, and the entity's level is determined, whether it is a core atom, an entity in the core atom's attributes, or both. Then, based on different judgment results, a search and match is performed in the database. Specifically, for the case where there is no subject but an entity, the search and match in the database based on different judgment results is as follows:

[0045] If only the core atom is located, the agricultural supply and demand information related to the matching is retrieved in the database according to the attributes of the core atom. Specifically, the technical attributes, consumption attributes and equipment attributes in the core atom attributes can be sequentially searched and matched. Among them, for the technical attributes of the core atom, priority is given to matching the manufacturers or studios that can provide planting and cultivation services for the core atom (i.e., the product). The purpose of adopting such a matching order is to attract agricultural supply and demand information demanders and to enable them to have an overall understanding of the product through the matched agricultural supply and demand information, because generally it is the agricultural supply and demand information demanders who need to obtain agricultural supply and demand information and perform search and matching. They may be interested in the planting and cultivation of the product and want to understand the current market situation.

[0046] If only entities in the core atomic attributes are located, the core atoms associated with the entities in the core atomic attributes are matched first, and then the core atomic attributes and the core atoms are matched with the associated agricultural supply and demand information in the database. Because the entities come from the attributes of the core atoms, the specific goals of the agricultural supply and demand information demanders cannot be determined. Therefore, the matching results can be sorted according to the release time and validity period of the agricultural supply and demand information, and then the sorted matching results can be provided to the agricultural supply and demand information demanders, which is conducive to the agricultural supply and demand information demanders obtaining satisfactory agricultural supply and demand information.

[0047] If the entities in the core atoms and core atomic attributes are located at the same time, the positions of all entities are first located in the knowledge graph, and the core atomic attributes therein are inferred through the knowledge graph. Then, the entities of the retrieved text are reorganized into new text according to the core atomic attributes, and the similarity between the new text and the relevant agricultural supply and demand information in the database is calculated and sorted, thereby providing the sorted matching results to the agricultural supply and demand information demanders, which is conducive to the agricultural supply and demand information demanders obtaining satisfactory agricultural supply and demand information.

[0048] It should be noted that the above-mentioned similarity calculation between the new text and the relevant agricultural supply and demand information in the database adopts the text similarity calculation technology in the natural language processing technology in the field of artificial intelligence, that is, by analyzing the two texts to calculate the similarity between the two texts.

[0049] For the second case, where there are both topics and entities, the extracted search text topic is used as the basis for search and matching. In the subsequent search and matching process, agricultural supply and demand information that is most closely related to the topic attributes of the search text is prioritized. Secondly, the entities in the extracted search text are placed in the knowledge graph for positioning, and the entity's level is determined, whether it is a core atom or an entity within the core atom's attributes. Then, based on different judgment results, the database is searched for matches. Specifically, for the case where there are both topics and entities, the database is searched for matches based on different judgment results as follows:

[0050] If only the core atom is located, the subject of the search text is used as the inference condition to match the relevant agricultural supply and demand information in the database, and the similarity between the matched agricultural supply and demand information and the search text is calculated and sorted. When sorting the agricultural supply and demand information, the validity period of the agricultural supply and demand information is added to the sorting calculation process, thereby providing the sorted matching results to the agricultural supply and demand information demanders, which is conducive to the agricultural supply and demand information demanders obtaining satisfactory agricultural supply and demand information;

[0051] If only the entity in the core atomic attribute is located, the attribute of the entity is matched with the topic of the search text. If the match is successful, it is associated with the core atom through the knowledge graph, and the associated core atom replaces the corresponding entity in the search text. The replaced new text is then similarly calculated and sorted with the matched agricultural supply and demand information. It can be understood that the entity based on the core atomic attribute may correspond to multiple core atoms, and the agricultural supply and demand information is time-sensitive. Therefore, when sorting the matching results, the release time of the agricultural supply and demand information can be used as the key condition for sorting, thereby providing the sorted matching results to the agricultural supply and demand information demanders, which is conducive to the agricultural supply and demand information demanders obtaining satisfactory agricultural supply and demand information.

[0052] For the third case, where there is no subject and no entity, keyword search matching is used during search and matching. The search text is matched with the agricultural supply and demand information in the database once, and then the detailed content and summary of the matched agricultural supply and demand information are similarly calculated with the search text for a second match. The matching results are sorted by similarity to filter out matching results with a similarity of less than 60%, which is conducive to those who need agricultural supply and demand information obtaining satisfactory agricultural supply and demand information.

[0053] It should be noted that the summary of the above-mentioned agricultural supply and demand information is obtained by using the summary extraction technology in the natural language processing technology in the field of artificial intelligence, that is, a summary is formed by calculating a piece of text, and the summary contains the main description content in the text.

[0054] On the other hand, this embodiment provides an agricultural supply and demand information matching system for implementing the above-mentioned agricultural supply and demand information matching method, and splits the agricultural supply and demand information matching system into functional modules, including a knowledge graph management module, a retrieval and matching module, and a supply and demand information entry module.

[0055] In this embodiment, the knowledge graph management module is used to construct a knowledge graph of agricultural supply and demand information in the database with products as core atoms, and assign attributes to the core atoms. The knowledge graph associates all agricultural supply and demand information related to the core atoms.

[0056] In this embodiment, the search and matching module is used to determine role attributes and establish matching rules. The search and matching module searches and matches relevant agricultural supply and demand information in the database using different matching rules according to the input search text, and displays the matching results.

[0057] In this embodiment, the supply and demand information entry module is used for agricultural supply and demand information publishers to enter agricultural supply and demand information into the database, and at the same time structure the data of agricultural supply and demand information and assign agricultural supply and demand information tags so that the agricultural supply and demand information can be associated with the corresponding core atoms.

[0058] It is understandable that the matching rules corresponding to different roles are different. Therefore, when entering agricultural supply and demand information, it is also necessary to first determine the role attributes to confirm whether the user is the publisher of agricultural supply and demand information. Subsequently, the publisher of agricultural supply and demand information can select the department to which the product in the entered agricultural supply and demand information belongs and the attribute information of the product, and finally fill in the title, detailed content, information validity period, address and other specific information of the agricultural supply and demand information. Among them, for the detailed content filled in, the supply and demand information entry module will use summary extraction technology to form the summary information of the agricultural supply and demand information.

[0059] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A method for matching agricultural supply and demand information, characterized in that: The following steps are involved: Step S1. Construct a knowledge graph of agricultural supply and demand information in the database with products as core atoms, and assign attributes to the core atoms. The knowledge graph associates all agricultural supply and demand information related to the core atoms; Step S2. Determine the role attributes and extract the subject and entity of the input search text respectively; Step S3. Establish matching rules, and use different matching rules to search for matches in the database to obtain matching results based on the role attributes and whether the subject and entity in the search text exist; In step S3, the matching rules include: If there is no subject but an entity, the entity of the extracted search text is placed in the knowledge graph for positioning, and the entity level is determined, whether it is a core atom, an entity in the core atom attribute, or both. Then, based on different judgment results, a match is searched in the database; There are topics and entities. The topic of the extracted search text is used as the basis. In the subsequent search and matching process, the agricultural supply and demand information that is most closely related to the topic attributes of the search text is prioritized. Secondly, the entities of the extracted search text are placed in the knowledge graph for positioning. The entity's level is determined, whether it is a core atom or an entity within the core atom's attributes. Then, based on different judgment results, the database is searched for matches. There is no subject and no entity. Keyword search and matching is used to search and match the search text with the agricultural supply and demand information in the database. The detailed content and summary of the searched agricultural supply and demand information are then calculated to perform a similarity calculation with the search text for a second search and match. The matching results are sorted by similarity to filter out matching results with a similarity of less than 60%.

2. The agricultural supply and demand information matching method according to claim 1, characterized in that: For the case where there is no subject but an entity, the search for matches in the database is as follows based on different judgment results: If only the core atom is located, the matching agricultural supply and demand information is retrieved from the database based on the attributes of the core atom; If only entities in the core atomic attributes are located, the core atoms associated with the entities in the core atomic attributes are retrieved and matched, and then the core atomic attributes and the core atoms are combined to retrieve the matching agricultural supply and demand information in the database, and the matching results are sorted according to the release time and validity period of the agricultural supply and demand information; if the core atoms and entities in the core atomic attributes are located at the same time, the positions of all entities are first located in the knowledge graph, and the core atomic attributes therein are inferred through the knowledge graph, and then the entities of the retrieved text are reorganized into new text according to the core atomic attributes, and the similarity between the new text and the relevant agricultural supply and demand information in the database is calculated and sorted.

3. The agricultural supply and demand information matching method according to claim 1, characterized in that: For the case where there are both subjects and entities, the specific search and matching in the database is as follows based on different judgment results: If only core atoms are located, the subject of the search text is used as the inference condition to search for relevant agricultural supply and demand information in the database. The similarity between the retrieved agricultural supply and demand information and the search text is calculated and sorted. When sorting the agricultural supply and demand information, the validity period of the agricultural supply and demand information is added to the sorting calculation process. If only entities in the core atomic attributes are located, the attributes of the entity are searched and matched with the subject of the search text. If the match is successful, it is associated with the core atom through the knowledge graph, and the associated core atom replaces the corresponding entity in the search text. The similarity between the replaced new text and the searched and matched agricultural supply and demand information is calculated and sorted.

4. The agricultural supply and demand information matching method according to claim 1, characterized in that: In step S1 , the attributes assigned to the core atom are divided into technical attributes, consumption attributes, equipment attributes, and information attributes.

5. The agricultural supply and demand information matching method according to claim 1, characterized in that: In step S2, the role attributes are divided into supply and demand information publisher and supply and demand information demander; Supply and demand information publishers can enter agricultural supply and demand information into the database. When entering agricultural supply and demand information, they can match the relevant information in the database according to the detailed content of the agricultural supply and demand information, and can regularly update and match the new agricultural supply and demand information released by other supply and demand information publishers; Supply and demand information demanders can enter search text to search for matching relevant agricultural supply and demand information in the database.

6. The agricultural supply and demand information matching method according to claim 1, characterized in that: In step S2, the subject of the retrieved text is extracted using a text subject classification technique, and the entities of the retrieved text are extracted using an entity recognition technique.

7. An agricultural supply and demand information matching system, used to implement the agricultural supply and demand information matching method according to any one of claims 1 to 6, characterized in that: The agricultural supply and demand information matching system is divided into functional modules, including: The knowledge graph management module is used to construct a knowledge graph of agricultural supply and demand information in the database with products as core atoms, and assign attributes to the core atoms. The knowledge graph will associate all agricultural supply and demand information related to the core atoms; The retrieval and matching module is used to determine role attributes and establish matching rules. The retrieval and matching module retrieves and matches relevant agricultural supply and demand information in the database using different matching rules based on the input search text and displays the matching results.

8. The agricultural supply and demand information matching system according to claim 7, characterized in that: The agricultural supply and demand information matching system also includes: The supply and demand information entry module is used to enter agricultural supply and demand information into the database, and at the same time structure the data of agricultural supply and demand information and assign agricultural supply and demand information labels.

Images