CN117807176A - Knowledge base index construction method, device and equipment based on two-dimensional gridding - Google Patents
Knowledge base index construction method, device and equipment based on two-dimensional gridding Download PDFInfo
- Publication number
- CN117807176A CN117807176A CN202311845980.0A CN202311845980A CN117807176A CN 117807176 A CN117807176 A CN 117807176A CN 202311845980 A CN202311845980 A CN 202311845980A CN 117807176 A CN117807176 A CN 117807176A
- Authority
- CN
- China
- Prior art keywords
- knowledge
- topic
- dimensional
- knowledge base
- map
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000010276 construction Methods 0.000 title claims abstract description 20
- 238000000034 method Methods 0.000 claims abstract description 22
- 238000013507 mapping Methods 0.000 claims description 17
- 230000004044 response Effects 0.000 claims description 15
- 238000012545 processing Methods 0.000 claims description 9
- 238000012216 screening Methods 0.000 claims description 5
- 239000000284 extract Substances 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000003993 interaction Effects 0.000 description 3
- 230000008685 targeting Effects 0.000 description 3
- 238000004590 computer program Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000009432 framing Methods 0.000 description 1
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention relates to the technical field of data retrieval, in particular to a knowledge base index construction method, device and equipment based on two-dimensional meshing, which comprises the following steps: extracting knowledge topics of the knowledge document, and carrying out coordinated abstraction on the knowledge topics of the knowledge document by utilizing a two-dimensional gridding method to obtain a topic index map taking two-dimensional coordinates as a retrieval index of the knowledge topics; in the topic index map, the topic knowledge base of each knowledge topic is linked to the corresponding two-dimensional coordinates through the corresponding relation between the two-dimensional coordinates and the knowledge topics, so as to obtain the knowledge base index map taking the two-dimensional coordinates as the retrieval index of the topic knowledge base. The invention constructs the knowledge base index, screens the documents according to the document vector index, rapidly determines the knowledge base documents required by replying to the retrieval requirement, and greatly improves the retrieval accuracy and the retrieval pertinence of the knowledge base.
Description
Technical Field
The invention relates to the technical field of data retrieval, in particular to a knowledge base retrieval method based on two-dimensional meshing.
Background
Knowledge base is a tool for the important knowledge storage, knowledge management and knowledge application of enterprises. The client can acquire knowledge that the client wants to consult and know by inputting the search words to search in the knowledge base.
In the prior art, after a client inputs a search term, the search term is matched with each knowledge in a knowledge base according to the text similarity of the search term and the knowledge in the knowledge base, and the matched knowledge related to the search term is fed back to the client as a search result. However, since knowledge base itself covers more knowledge, the search range is large, and the search result based on similarity is very much, so that many knowledge that is not needed by the client itself is covered, so that it is difficult for the client to quickly find the knowledge needed by the client from the fed-back large amount of knowledge. Therefore, the probability that the search result obtained by the existing search method contains the knowledge required by the target client is low, and the search experience brought to the client is poor.
Disclosure of Invention
The invention aims to provide a two-dimensional gridding-based knowledge base retrieval method, device and equipment, which are used for solving the technical problems that in the prior art, the probability of containing knowledge required by a target client in a retrieval result is low and the retrieval experience brought to the client is poor.
In order to solve the technical problems, the invention specifically provides the following technical scheme:
in a first aspect of the present invention, a two-dimensional gridding-based knowledge base index construction method includes the steps of:
acquiring a knowledge base, wherein the knowledge base comprises a plurality of knowledge documents;
extracting knowledge topics of the knowledge document, and carrying out coordinated abstraction on the knowledge topics of the knowledge document by utilizing a two-dimensional gridding method to obtain a topic index map taking two-dimensional coordinates as a retrieval index of the knowledge topics;
acquiring all knowledge documents of each knowledge topic, and combining the knowledge documents into a topic knowledge base of each knowledge topic;
in the topic index map, a topic knowledge base of each knowledge topic is linked to the corresponding two-dimensional coordinates through the corresponding relation between the two-dimensional coordinates and the knowledge topics, so as to obtain a knowledge base index map taking the two-dimensional coordinates as the retrieval index of the topic knowledge base;
carrying out document vectorization processing on a topic knowledge base linked with a retrieval index in a selected state in a knowledge base index map by using a Langchain framework to obtain a Langchain vector data set for providing effective information for generating retrieval replies;
and carrying out knowledge summarization based on the retrieval requirements by combining the Langcain vector data set with a large language model LLM, and generating a retrieval response meeting the retrieval requirements.
As a preferred solution of the present invention, the building of the topic index map includes:
setting two-dimensional coordinates for each knowledge topic one by one;
and carrying out two-dimensional gridding mapping on each knowledge topic according to the two-dimensional coordinates to obtain the topic index map.
As a preferred embodiment of the present invention, the setting of the two-dimensional coordinates includes:
acquiring WebGis geographic map coordinates of each knowledge topic;
the WebGis geographic map coordinates of the knowledge base are set as two-dimensional coordinates of the knowledge base.
As a preferable scheme of the invention, the WebGis geographic map coordinates of the knowledge topics are mapped into the WebGis geographic map through two-dimensional meshing, so as to obtain a topic index map.
As a preferred embodiment of the present invention, the setting of the two-dimensional coordinates includes:
acquiring association relations on a non-WebGis geographic map among the knowledge topics;
constructing a relationship map among knowledge topics based on the association relationship on the non-WebGis geographic map among the knowledge topics;
acquiring coordinates of each knowledge topic in a relationship map;
the coordinates of the knowledge topics in the relationship map are set as two-dimensional coordinates.
As a preferable scheme of the invention, the coordinates of the knowledge topics in the relation map are subjected to two-dimensional gridding mapping to the relation map, so as to obtain the topic index map.
As a preferred scheme of the present invention, the association relationship includes a two-dimensional low-latitude association relationship and a two-dimensional high-latitude association relationship, the relationship map includes a two-dimensional low-latitude relationship map and a two-dimensional high-latitude relationship map, wherein,
the two-dimensional coordinates of the knowledge topics in the two-dimensional low-latitude relation map are subjected to two-dimensional gridding mapping to the two-dimensional low-latitude relation map, and a topic index map is obtained;
and performing two-dimensional gridding mapping on the two-dimensional coordinates of the knowledge topics in the two-dimensional high-latitude relation map to obtain a topic index map.
As a preferred embodiment of the present invention, the generating of the LangChain vector dataset includes: extracting keywords from the search requirements;
searching according to similarity matching in a topic knowledge base by utilizing the keywords of the search requirement to obtain knowledge points related to the search requirement in the topic knowledge base;
and taking knowledge points related to the retrieval requirements in the topic knowledge base as effective information for generating retrieval answers to form a Langchain vector data set.
In a second aspect of the present invention, a two-dimensional gridding-based knowledge base index construction apparatus is characterized by comprising:
the data acquisition unit acquires a knowledge base and a search requirement, wherein the knowledge base comprises a plurality of knowledge documents;
the index construction unit extracts the knowledge topics of the knowledge document, and performs coordinated abstraction on the knowledge topics of the knowledge document by using a two-dimensional gridding method to obtain a topic index map taking the two-dimensional coordinates as a retrieval index of the knowledge topics;
acquiring all knowledge documents of each knowledge topic, and combining the knowledge documents into a topic knowledge base of each knowledge topic;
in the topic index map, a topic knowledge base of each knowledge topic is linked to the corresponding two-dimensional coordinates through the corresponding relation between the two-dimensional coordinates and the knowledge topics, so as to obtain a knowledge base index map taking the two-dimensional coordinates as the retrieval index of the topic knowledge base;
the knowledge base screening unit is used for carrying out document vectorization processing on a topic knowledge base linked with a retrieval index in a selected state in a knowledge base index map by utilizing a Langchain framework to obtain a Langchain vector data set for providing effective information for generating a retrieval response;
and the retrieval response unit is used for summarizing knowledge based on retrieval requirements by combining the Langcain vector data set with a large language model LLM and generating retrieval responses meeting the retrieval requirements.
In a third aspect of the present invention, a computer readable medium is provided, on which a computer program is stored, wherein the program when executed by a processor implements a two-dimensional gridding based knowledge base retrieval method.
In a fourth aspect of the invention, the invention provides a computer device comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor;
wherein the memory stores instructions executable by the at least one processor to cause the computer device to perform a two-dimensional gridding-based knowledge base retrieval method.
Compared with the prior art, the invention has the following beneficial effects:
the invention builds the knowledge base index, screens the documents according to the document vector index, rapidly determines the knowledge base documents required by replying to the retrieval requirement, greatly improves the retrieval accuracy of the knowledge base, and combines the Langchain framework and the large language model LLM to carry out retrieval reply, thereby improving the man-machine interaction of the retrieval, leading the retrieval reply to more accord with the retrieval requirement, and having higher efficiency and higher readability.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below. It will be apparent to those of ordinary skill in the art that the drawings in the following description are exemplary only and that other implementations can be obtained from the extensions of the drawings provided without inventive effort.
FIG. 1 is a flowchart of a two-dimensional gridding-based knowledge base retrieval method according to an embodiment of the present invention;
FIG. 2 is a block diagram of a knowledge base retrieval device based on two-dimensional meshing provided by an embodiment of the invention;
FIG. 3 is a diagram illustrating an internal structure of a computer device according to an embodiment of the present invention;
FIG. 4 is a knowledge base index map using a WebGis geographic map as a carrier, according to an embodiment of the present invention;
fig. 5 is a schematic diagram of a search reply according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
As shown in fig. 1, in a first aspect of the present invention, the present invention provides a two-dimensional gridding-based knowledge base index construction method, which is characterized by comprising the following steps:
acquiring a knowledge base, wherein the knowledge base comprises a plurality of knowledge documents;
extracting knowledge topics of the knowledge document, and carrying out coordinated abstraction on the knowledge topics of the knowledge document by utilizing a two-dimensional gridding method to obtain a topic index map taking two-dimensional coordinates as a retrieval index of the knowledge topics;
acquiring all knowledge documents of each knowledge topic, and combining the knowledge documents into a topic knowledge base of each knowledge topic;
in the topic index map, a topic knowledge base of each knowledge topic is linked to the corresponding two-dimensional coordinates through the corresponding relation between the two-dimensional coordinates and the knowledge topics, so as to obtain a knowledge base index map taking the two-dimensional coordinates as the retrieval index of the topic knowledge base;
carrying out document vectorization processing on a topic knowledge base linked with a retrieval index in a selected state in a knowledge base index map by using a Langchain framework to obtain a Langchain vector data set for providing effective information for generating retrieval replies;
and carrying out knowledge summarization based on the retrieval requirements by combining the Langcain vector data set with a large language model LLM, and generating a retrieval response meeting the retrieval requirements.
In order to improve the searching pertinence, the method and the device can quickly position the knowledge document containing the related knowledge content required by searching by constructing the knowledge base index to carry out the pertinence screening of the searching range, avoid invalid searching, obtain a large number of invalid contents and optimize the searching experience of users.
According to the invention, knowledge retrieval is performed in a targeted database by using the Langchain framework and combining with the large language model LLM, so that the efficiency of retrieving effective knowledge information meeting the retrieval requirement is higher, the accuracy is higher, and the information is processed into retrieval answers with higher readability, thereby facilitating the user to grasp retrieval knowledge points and further optimizing the retrieval experience of the user.
Specifically, the invention endows the knowledge base with graphical retrieval indexes, carries out coordinate selection on the knowledge base from a graphical interface, determines the knowledge base for retrieval in a coordinate selection mode, reduces the retrieval range from all knowledge documents of the whole knowledge base to knowledge documents of part of the knowledge base, effectively compresses the retrieval range, improves the retrieval pertinence and the accuracy, and comprises the following specific steps:
the construction of the topic index map comprises the following steps:
setting two-dimensional coordinates for each knowledge topic one by one;
and carrying out two-dimensional gridding mapping on each knowledge topic according to the two-dimensional coordinates to obtain a topic index map.
The setting of two-dimensional coordinates includes:
acquiring WebGis geographic map coordinates of each knowledge topic;
the WebGis geographic map coordinates of the knowledge base are set as two-dimensional coordinates of the knowledge base.
And performing two-dimensional gridding mapping on the WebGis geographic map coordinates of the knowledge theme to obtain a theme index map.
The setting of two-dimensional coordinates includes:
acquiring association relations on a non-WebGis geographic map among all knowledge topics;
constructing a relationship map among knowledge topics based on the association relationship on the non-WebGis geographic map among the knowledge topics;
acquiring coordinates of each knowledge topic in a relation map;
the coordinates of the knowledge topics in the relationship map are set as two-dimensional coordinates.
And carrying out two-dimensional gridding mapping on the coordinates of the knowledge topics in the relation map to obtain a topic index map.
The association relationship comprises a two-dimensional low-latitude association relationship and a two-dimensional high-latitude association relationship, the relationship map comprises a two-dimensional low-latitude relationship map and a two-dimensional high-latitude relationship map, wherein,
performing two-dimensional gridding mapping on the two-dimensional coordinates of the knowledge topics in the two-dimensional low-latitude relation map to obtain a topic index map;
and carrying out two-dimensional gridding mapping on the two-dimensional coordinates of the knowledge topics in the two-dimensional low-latitude relation map to obtain a topic index map.
The generation of LangChain vector data sets comprises: extracting keywords from the search requirements;
searching according to similarity matching in a topic knowledge base by utilizing the keywords of the search requirement to obtain knowledge points related to the search requirement in the topic knowledge base;
and taking knowledge points related to the retrieval requirements in the topic knowledge base as effective information for generating retrieval answers to form a Langchain vector data set.
As shown in fig. 4 to 5, the first embodiment: acquiring a knowledge base for geological knowledge in the geological industry, wherein the knowledge base comprises a plurality of knowledge documents;
extracting geographic positions of the knowledge documents, and obtaining WebGis geographic map coordinates of each geographic position;
the WebGis geographic map coordinates of the geographic location are set as two-dimensional coordinates of the geographic location.
And performing two-dimensional gridding mapping on the WebGis geographic map coordinates of the geographic position to obtain a theme index map.
Acquiring all knowledge documents of each geographic position, and combining the knowledge documents into a geological knowledge base of each geographic position;
in the topic index map, the geological knowledge base of each geographic position is linked to the corresponding two-dimensional coordinates through the corresponding relation between the two-dimensional coordinates and the geographic positions to obtain a knowledge base index map taking the two-dimensional coordinates as the index of the topic knowledge base, and the carrier is a WebGis geographic map, namely, the carrier is presented in the form of the WebGis geographic map.
The relevant knowledge base of the geographic position can be obtained by selecting one geographic position in the WebGis geographic map, the relevant knowledge base can be used for retrieving the geological information for knowing the geographic position, the information of the geographic position can be directly indexed into the knowledge base for targeted retrieval when being retrieved, invalid searching of other additional knowledge bases can not be generated, and the targeting efficiency is high.
Carrying out document vectorization processing on a topic knowledge base linked with a retrieval index in a selected state in a knowledge base index map by using a Langchain framework to obtain a Langchain vector data set for providing effective information for generating retrieval replies;
and carrying out knowledge summarization based on the retrieval requirements by combining the Langcain vector data set with a large language model LLM, and generating a retrieval response meeting the retrieval requirements.
The coordinate index may be further generalized and is not limited to operating only on a geographic map, as in the second embodiment.
Second embodiment: acquiring a two-dimensional low-latitude association relationship on a non-WebGis geographic map among geographic positions;
based on the two-dimensional low-latitude association relationship on the non-WebGis geographic map among the geographic positions, a two-dimensional low-latitude relationship map among the geographic positions is constructed, and a triangle network can be used for linking the geographic positions to form a relationship map without adopting the spatial relationship among the geographic positions on the geographic map;
acquiring coordinates of each geographic position in a two-dimensional low-latitude relation map;
and setting the coordinates of the geographic position in the two-dimensional low-latitude relation map as two-dimensional coordinates.
And carrying out two-dimensional gridding mapping on the coordinates of the geographic position in the two-dimensional low-latitude relation map to obtain the theme index map.
Acquiring all knowledge documents of each geographic position, and combining the knowledge documents into a geological knowledge base of each geographic position;
in the topic index map, the geological knowledge base of each geographical position is linked to the corresponding two-dimensional coordinates through the corresponding relation between the two-dimensional coordinates and the geographical positions to obtain a knowledge base index map taking the two-dimensional coordinates as the retrieval index of the topic knowledge base, and the carrier is a two-dimensional low-latitude relation map, namely, the carrier is presented in the form of the two-dimensional low-latitude relation map.
The method has the advantages that a geographic position is selected in the two-dimensional low-latitude relation map, so that a related knowledge base of the geographic position can be obtained, geological information for knowing the geographic position can be searched, the information of the geographic position can be directly indexed into the knowledge base for targeted search when being searched, invalid searching of other additional knowledge bases can not be generated, and the method is high in targeting efficiency.
Carrying out document vectorization processing on a topic knowledge base linked with a retrieval index in a selected state in a knowledge base index map by using a Langchain framework to obtain a Langchain vector data set for providing effective information for generating retrieval replies;
and carrying out knowledge summarization based on the retrieval requirements by combining the Langcain vector data set with a large language model LLM, and generating a retrieval response meeting the retrieval requirements.
Further, the relationship map can be generalized to a two-dimensional high latitude level, as in the third embodiment.
Third embodiment: acquiring a two-dimensional high-latitude association relationship on a non-WebGis geographic map among various geographic positions;
based on the two-dimensional high-latitude association relationship on the non-WebGis geographic map among the geographic positions, a two-dimensional high-latitude relationship map among the geographic positions is constructed, and a triangle network can be used for linking the geographic positions to form a relationship map without adopting the spatial relationship among the geographic positions on the geographic map;
acquiring coordinates of each geographic position in a two-dimensional high-latitude relation map;
and setting the coordinates of the geographic position in the two-dimensional high-latitude relation map as two-dimensional coordinates.
And carrying out two-dimensional gridding mapping on the coordinates of the geographic position in the two-dimensional high-latitude relation map to obtain the theme index map.
Acquiring all knowledge documents of each geographic position, and combining the knowledge documents into a geological knowledge base of each geographic position;
in the topic index map, the geological knowledge base of each geographical position is linked to the corresponding two-dimensional coordinates through the corresponding relation between the two-dimensional coordinates and the geographical positions to obtain a knowledge base index map taking the two-dimensional coordinates as the retrieval index of the topic knowledge base, and the carrier is a two-dimensional high-latitude relation map, namely, the two-dimensional high-latitude relation map is presented.
The method has the advantages that a geographic position is selected in the two-dimensional high-latitude relation map, so that a related knowledge base of the geographic position can be obtained, geological information for knowing the geographic position can be searched, the information of the geographic position can be directly indexed into the knowledge base for targeted search when being searched, invalid searching of other additional knowledge bases can not be generated, and the method is high in targeting efficiency.
Carrying out document vectorization processing on a topic knowledge base linked with a retrieval index in a selected state in a knowledge base index map by using a Langchain framework to obtain a Langchain vector data set for providing effective information for generating retrieval replies;
and carrying out knowledge summarization based on the retrieval requirements by combining the Langcain vector data set with a large language model LLM, and generating a retrieval response meeting the retrieval requirements.
Therefore, the invention adds a new text interaction mode, and realizes the screening index of the data through the operations of selecting and framing various graphs (geographic map, low latitude relation map and high latitude relation map).
In a second aspect of the present invention, a two-dimensional gridding-based knowledge base index construction apparatus is characterized by comprising:
the data acquisition unit acquires a knowledge base and a search requirement, wherein the knowledge base comprises a plurality of knowledge documents;
the index construction unit extracts the knowledge topics of the knowledge document, and performs coordinated abstraction on the knowledge topics of the knowledge document by using a two-dimensional gridding method to obtain a topic index map taking the two-dimensional coordinates as a retrieval index of the knowledge topics;
acquiring all knowledge documents of each knowledge topic, and combining the knowledge documents into a topic knowledge base of each knowledge topic;
in the topic index map, a topic knowledge base of each knowledge topic is linked to the corresponding two-dimensional coordinates through the corresponding relation between the two-dimensional coordinates and the knowledge topics, so as to obtain a knowledge base index map taking the two-dimensional coordinates as the retrieval index of the topic knowledge base;
the knowledge base screening unit is used for carrying out document vectorization processing on a topic knowledge base linked with a retrieval index in a selected state in a knowledge base index map by utilizing a Langchain framework to obtain a Langchain vector data set for providing effective information for generating a retrieval response;
and the retrieval response unit is used for summarizing knowledge based on retrieval requirements by combining the Langcain vector data set with a large language model LLM and generating retrieval responses meeting the retrieval requirements.
In a third aspect of the present invention, a computer readable medium is provided, on which a computer program is stored, wherein the program when executed by a processor implements a two-dimensional gridding based knowledge base retrieval method.
In a fourth aspect of the invention, the invention provides a computer device comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor;
wherein the memory stores instructions executable by the at least one processor to cause the computer device to perform a two-dimensional gridding-based knowledge base retrieval method.
The invention builds the knowledge base index, screens the documents according to the document vector index, rapidly determines the knowledge base documents required by replying to the retrieval requirement, greatly improves the retrieval accuracy of the knowledge base, and combines the Langchain framework and the large language model LLM to carry out retrieval reply, thereby improving the man-machine interaction of the retrieval, leading the retrieval reply to more accord with the retrieval requirement, and having higher efficiency and higher readability.
The above embodiments are only exemplary embodiments of the present application and are not intended to limit the present application, the scope of which is defined by the claims. Various modifications and equivalent arrangements may be made to the present application by those skilled in the art, which modifications and equivalents are also considered to be within the scope of the present application.
Claims (10)
1. The knowledge base index construction method based on two-dimensional gridding is characterized by comprising the following steps of:
acquiring a knowledge base, wherein the knowledge base comprises a plurality of knowledge documents;
extracting knowledge topics of the knowledge document, and carrying out coordinated abstraction on the knowledge topics of the knowledge document by utilizing a two-dimensional gridding method to obtain a topic index map taking two-dimensional coordinates as a retrieval index of the knowledge topics;
acquiring all knowledge documents of each knowledge topic, and combining the knowledge documents into a topic knowledge base of each knowledge topic;
in the topic index map, a topic knowledge base of each knowledge topic is linked to the corresponding two-dimensional coordinates through the corresponding relation between the two-dimensional coordinates and the knowledge topics, so as to obtain a knowledge base index map taking the two-dimensional coordinates as the retrieval index of the topic knowledge base;
carrying out document vectorization processing on a topic knowledge base linked with a retrieval index in a selected state in a knowledge base index map by using a Langchain framework to obtain a Langchain vector data set for providing effective information for generating retrieval replies;
and carrying out knowledge summarization based on the retrieval requirements by combining the Langcain vector data set with a large language model LLM, and generating a retrieval response meeting the retrieval requirements.
2. The two-dimensional gridding-based knowledge base index construction method according to claim 1, characterized by comprising the steps of: the construction of the topic index map comprises the following steps:
setting two-dimensional coordinates for each knowledge topic one by one;
and carrying out two-dimensional gridding mapping on each knowledge topic according to the two-dimensional coordinates to obtain the topic index map.
3. The two-dimensional gridding-based knowledge base index construction method according to claim 1, characterized by comprising the steps of: the setting of the two-dimensional coordinates comprises the following steps:
acquiring WebGis geographic map coordinates of each knowledge topic;
the WebGis geographic map coordinates of the knowledge base are set as two-dimensional coordinates of the knowledge base.
4. The two-dimensional gridding-based knowledge base retrieval method according to claim 1, wherein: and performing two-dimensional gridding mapping on the WebGis geographic map coordinates of the knowledge topics to obtain a topic index map.
5. The two-dimensional gridding-based knowledge base index construction method according to claim 1, characterized by comprising the steps of: the setting of the two-dimensional coordinates comprises the following steps:
acquiring association relations on a non-WebGis geographic map among the knowledge topics;
constructing a relationship map among knowledge topics based on the association relationship on the non-WebGis geographic map among the knowledge topics;
acquiring coordinates of each knowledge topic in a relationship map;
the coordinates of the knowledge topics in the relationship map are set as two-dimensional coordinates.
6. The two-dimensional gridding-based knowledge base index construction method according to claim 5, wherein: and carrying out two-dimensional gridding mapping on the coordinates of the knowledge topics in the relation map to obtain a topic index map.
7. The two-dimensional gridding-based knowledge base index construction method according to claim 6, wherein: the association relationship comprises a two-dimensional low-latitude association relationship and a two-dimensional high-latitude association relationship, the relationship map comprises a two-dimensional low-latitude relationship map and a two-dimensional high-latitude relationship map, wherein,
the two-dimensional coordinates of the knowledge topics in the two-dimensional low-latitude relation map are subjected to two-dimensional gridding mapping to the two-dimensional low-latitude relation map, and a topic index map is obtained;
and performing two-dimensional gridding mapping on the two-dimensional coordinates of the knowledge topics in the two-dimensional high-latitude relation map to obtain a topic index map.
8. The two-dimensional gridding-based knowledge base index construction method according to claim 1, characterized by comprising the steps of: the generating of the LangChain vector data set comprises the following steps: extracting keywords from the search requirements;
searching according to similarity matching in a topic knowledge base by utilizing the keywords of the search requirement to obtain knowledge points related to the search requirement in the topic knowledge base;
and taking knowledge points related to the retrieval requirements in the topic knowledge base as effective information for generating retrieval answers to form a Langchain vector data set.
9. A two-dimensional gridding-based knowledge base index construction apparatus, comprising:
the data acquisition unit acquires a knowledge base and a search requirement, wherein the knowledge base comprises a plurality of knowledge documents;
the index construction unit extracts the knowledge topics of the knowledge document, and performs coordinated abstraction on the knowledge topics of the knowledge document by using a two-dimensional gridding method to obtain a topic index map taking the two-dimensional coordinates as a retrieval index of the knowledge topics;
acquiring all knowledge documents of each knowledge topic, and combining the knowledge documents into a topic knowledge base of each knowledge topic;
in the topic index map, a topic knowledge base of each knowledge topic is linked to the corresponding two-dimensional coordinates through the corresponding relation between the two-dimensional coordinates and the knowledge topics, so as to obtain a knowledge base index map taking the two-dimensional coordinates as the retrieval index of the topic knowledge base;
the knowledge base screening unit is used for carrying out document vectorization processing on a topic knowledge base linked with a retrieval index in a selected state in a knowledge base index map by utilizing a Langchain framework to obtain a Langchain vector data set for providing effective information for generating a retrieval response;
and the retrieval response unit is used for summarizing knowledge based on retrieval requirements by combining the Langcain vector data set with a large language model LLM and generating retrieval responses meeting the retrieval requirements.
10. A computer device, comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor;
wherein the memory stores instructions executable by the at least one processor to cause a computer device to perform the method of any of claims 1-8.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311845980.0A CN117807176A (en) | 2023-12-29 | 2023-12-29 | Knowledge base index construction method, device and equipment based on two-dimensional gridding |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311845980.0A CN117807176A (en) | 2023-12-29 | 2023-12-29 | Knowledge base index construction method, device and equipment based on two-dimensional gridding |
Publications (1)
Publication Number | Publication Date |
---|---|
CN117807176A true CN117807176A (en) | 2024-04-02 |
Family
ID=90429352
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202311845980.0A Pending CN117807176A (en) | 2023-12-29 | 2023-12-29 | Knowledge base index construction method, device and equipment based on two-dimensional gridding |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN117807176A (en) |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102096713A (en) * | 2011-01-29 | 2011-06-15 | 广州都市圈网络科技有限公司 | Grid-based two-dimensional or three-dimensional map matching method and system |
US20120324347A1 (en) * | 2011-06-15 | 2012-12-20 | Terrence Monroe | Method for parsing, searching and formatting of text input for visual mapping of knowledge information |
KR20190041788A (en) * | 2017-10-13 | 2019-04-23 | 네이버랩스 주식회사 | Method and system for providing user location information using gridding map |
CN114328766A (en) * | 2020-09-27 | 2022-04-12 | 中国科学院空天信息创新研究院 | Method, device, medium and equipment for constructing geographic knowledge database |
CN116842126A (en) * | 2023-08-29 | 2023-10-03 | 青岛网信信息科技有限公司 | Method, medium and system for realizing accurate output of knowledge base by using LLM |
CN117033579A (en) * | 2023-07-14 | 2023-11-10 | 北银金融科技有限责任公司 | Novel knowledge base system and method based on LLMS |
-
2023
- 2023-12-29 CN CN202311845980.0A patent/CN117807176A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102096713A (en) * | 2011-01-29 | 2011-06-15 | 广州都市圈网络科技有限公司 | Grid-based two-dimensional or three-dimensional map matching method and system |
US20120324347A1 (en) * | 2011-06-15 | 2012-12-20 | Terrence Monroe | Method for parsing, searching and formatting of text input for visual mapping of knowledge information |
KR20190041788A (en) * | 2017-10-13 | 2019-04-23 | 네이버랩스 주식회사 | Method and system for providing user location information using gridding map |
CN114328766A (en) * | 2020-09-27 | 2022-04-12 | 中国科学院空天信息创新研究院 | Method, device, medium and equipment for constructing geographic knowledge database |
CN117033579A (en) * | 2023-07-14 | 2023-11-10 | 北银金融科技有限责任公司 | Novel knowledge base system and method based on LLMS |
CN116842126A (en) * | 2023-08-29 | 2023-10-03 | 青岛网信信息科技有限公司 | Method, medium and system for realizing accurate output of knowledge base by using LLM |
Non-Patent Citations (2)
Title |
---|
徐照;张路;索华;迟英姿;: "基于工业基础类的建筑物3D Tiles数据可视化", 浙江大学学报(工学版), no. 06, 12 April 2019 (2019-04-12) * |
黄玉青;建明;赵雅鹏;张号;: "基于WebGL的三维WebGIS应用研究", 测绘技术装备, no. 02, 25 June 2017 (2017-06-25), pages 33 - 36 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US7444343B2 (en) | Hybrid location and keyword index | |
US11899681B2 (en) | Knowledge graph building method, electronic apparatus and non-transitory computer readable storage medium | |
CN111694965B (en) | Image scene retrieval system and method based on multi-mode knowledge graph | |
US10452661B2 (en) | Automated database schema annotation | |
US20140122535A1 (en) | Extracting Semantic Relationships from Table Structures in Electronic Documents | |
CN101350013A (en) | Method and system for searching geographical information | |
CN106202514A (en) | Accident based on Agent is across the search method of media information and system | |
CN108228743A (en) | A kind of real-time big data search engine system | |
CN111428047A (en) | Knowledge graph construction method and device based on UC L semantic indexing | |
US11977567B2 (en) | Method of retrieving query, electronic device and medium | |
CN113342976A (en) | Method, device, storage medium and equipment for automatically acquiring and processing data | |
US8700624B1 (en) | Collaborative search apps platform for web search | |
CN109783484A (en) | The construction method and system of the data service platform of knowledge based map | |
CN105608113A (en) | Method and apparatus for judging POI data in text | |
US9323833B2 (en) | Relevant online search for long queries | |
CA2651284C (en) | Coordinate system identification | |
CN111933141A (en) | Artificial intelligence voice interaction system based on big data | |
CN111930891A (en) | Retrieval text expansion method based on knowledge graph and related device | |
CN117807176A (en) | Knowledge base index construction method, device and equipment based on two-dimensional gridding | |
CN106649883B (en) | cross-language theme website automatic discovery method | |
CN108959555B (en) | Query type expansion method and device, computer equipment and storage medium | |
CN113076395A (en) | Semantic model training, search and display method, device, equipment and storage medium | |
CN105159899A (en) | Searching method and searching device | |
Ladra et al. | A toponym resolution service following the OGC WPS standard | |
CN105550285A (en) | Method and device for building POI dictionary |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |