CN-121996853-A - Entity space-time information and topological relation retrieval method based on space-time knowledge graph

CN121996853ACN 121996853 ACN121996853 ACN 121996853ACN-121996853-A

Abstract

The invention discloses a space-time knowledge graph-based entity space-time information and topological relation retrieval method, which comprises the steps of constructing a knowledge source database, carrying out space knowledge graph processing to obtain a space-time knowledge graph database, carrying out semantic segmentation on input query data by using a large language model LLM, carrying out fact statement sentence extraction on each segmentation, extracting query combinations comprising triples, time entities and space geography entities from the fact statement sentences, creating a multi-path parallel retrieval path matched with the query combinations by using a space-time retrieval planner, carrying out multi-path parallel search in the space-time knowledge graph database by using the multi-path parallel retrieval path, and summarizing and layering to output retrieval results, wherein the multi-path parallel retrieval path comprises triples retrieval, time retrieval and space geography retrieval. The invention improves the precision and generalization of complex space-time semantic understanding and realizes multidimensional collaborative retrieval optimization.

Inventors

ZHANG BO
PENG XIAOBO
MA XIAOFEI
DAI XUQIANG
Qiao Hanyang

Assignees

浙江时空智子大数据有限公司

Dates

Publication Date: 20260508
Application Date: 20260126

Claims (10)

1. A retrieval method of entity space-time information and topological relation based on space-time knowledge graph is characterized by comprising the following steps: S1, constructing a knowledge source database, and performing space-time knowledge graph processing to obtain a space-time knowledge graph database, wherein the space-time knowledge graph comprises a time entity, a space geographic entity and a triplet; S2, carrying out semantic segmentation on input query data by utilizing a large language model LLM, extracting a fact statement sentence from each segmentation, and extracting a query combination comprising a triplet, a time entity and a space geographic entity from the fact statement sentence; S3, creating a plurality of paths of parallel search paths matched with the query combination by using a space-time search planner, wherein the plurality of paths of parallel search comprise triplet search, time search and space geographic search, and performing multi-path parallel search in a space-time knowledge graph database by using the plurality of paths of parallel search, and summarizing and layering to output search results.
2. The method for searching entity space-time information and topological relation based on space-time knowledge graph according to claim 1, wherein in the method S1, the method for obtaining the space-time knowledge graph database is as follows: S11, carrying out semantic segmentation on data in a knowledge source database by utilizing a large language model LLM, extracting a fact statement sentence from each semantic segmentation, and respectively extracting a triplet, a time entity and a space geographic entity from the fact statement sentence, wherein the triplet comprises an entity and an entity relation; s12, performing time attribute coding on the time entity, constructing a gridded Beidou satellite geographic map, wherein all grids in the Beidou satellite geographic map correspond to geographic codes to form a Beidou grid coding dataset, and performing associated grid matching coding on the space entity by using the Beidou grid coding dataset to form corresponding space attributes; s13, associating and collecting the time entity containing the time attribute and the space geographic entity containing the space attribute in the knowledge source database with the triples, and constructing to obtain a space-time knowledge graph database.
3. The method S3 is characterized in that the triplet search utilizes the triplet in the query combination to create an entity search, the relation search and clustering search parallel search task, the entity search task uses the entity in the triplet in the query combination as a keyword to perform keyword search, the relation search task uses the entity relation in the triplet in the query combination to perform vector representation to perform similarity search, the clustering search task uses the entity in the triplet in the query combination to perform similar entity clustering and search according to the similar entity clustering result, the time search utilizes the time entity in the query combination to create a time search task and a time reasoning search task, the time search task uses the time entity in the query combination as a keyword to perform keyword search, the time reasoning search task uses the time entity in the query combination to perform inference calculation of absolute time, relative time and time to perform search, the space geographic search uses the space geographic entity in the query combination to create a space search, the space attribute search task of the space geographic entity in the query combination to perform space attribute search, and the space attribute search task uses the space geographic entity in the query combination to perform space attribute search of the entity.
4. The method for searching entity space-time information and topological relation based on the space-time knowledge graph is characterized by further comprising a multi-hop path searching task, wherein the multi-hop path searching task is used for searching a plurality of entity hop neighborhoods by taking an entity of a triplet in a query combination as a center, the time searching further comprises a time multi-hop searching task, the time multi-hop searching task is used for searching a plurality of time entity multi-hop neighborhoods by taking a time entity in the query combination as a center, the space geographic searching further comprises a space multi-hop searching task, the space multi-hop searching task is used for searching a plurality of space geographic entity multi-hop neighborhoods by taking a space geographic entity in the query combination as a center, the triple searching, the time searching and the space geographic searching are respectively and parallelly executed, the searching task is marked according to entity similarity setting layering weight parameters in the query combination, and search results respectively output by the triple searching, the time searching and the space geographic searching are summarized and layered.
5. The method for searching the entity space-time information and the topological relation based on the space-time knowledge graph is characterized by further comprising a time event analysis module and a time trend analysis module, wherein the time event analysis module is used for carrying out same or/and similar entity association analysis on search results output by the time search, the time trend analysis module is used for carrying out trend analysis of the same entity or/and trend comparison analysis of similar entities on the search results output by the time search, the space geographic search further comprises a space attribute particle level search task and a space attribute coding topological search task, the space attribute particle level search task uses the space attribute of the space geographic entity as the center, the space attribute coding topological search task uses the space attribute of the space geographic entity as the center, the space topological relation comprises a spatial position relation which comprises the space geographic entity as a reference, intersection, adjacency and relative orientation, the space geographic search further comprises a space event analysis module and a space trend analysis module, and the space event analysis module is used for carrying out same or/and/or similar entity association analysis on the search results output by the space event analysis module, and the space trend analysis module is used for carrying out same or/and comparison analysis on the search results output by the space trend analysis of the space entity.
6. The method for searching for entity spatiotemporal information and topological relation based on spatiotemporal knowledge graph according to claim 2, characterized in that in method S11, the statement extraction is performed on the blocks, and the method is adopted from the blocks Function extraction and atomization fact statement sentence set The expression is as follows: Wherein In the case of a large language model, For the purpose of semantic blocking, In order to extract the prompt words, in the method S13, time entities containing time attributes are associated in time sequence according to time attribute codes, and space geographic entities containing space attributes are associated in space based on Beidou satellite geographic maps.
7. The method for searching for entity spatiotemporal information and topological relation based on spatiotemporal knowledge graph according to claim 2, characterized in that in method S12, the time entity The encoded temporal attribute expression is as follows: , Is an entity The unique identification is provided with a unique identification, Representing the description information of the entity, As a matter of time-type, 、 The spatial attribute expressions of the spatial geographic entities are as follows: Wherein Is an entity The unique identification is provided with a unique identification, Representing the description information of the entity, In the form of a spatial geographic type, Corresponding the longitude and latitude and the altitude of the geographic coordinates for the entity; And And respectively representing the Beidou grid code set and the Beidou grid level corresponding to the space geographic entity.
8. The method for searching for entity space-time information and topological relation based on space-time knowledge graph of claim 2, wherein the method S12 further comprises the following steps: s121, pass through Function acquisition of spatially or geographically correlated sets of spatial geographic locations By means of Function judgment and acquisition of space geographic position The spatial geographic type and Beidou grid level of (C) are obtained Corresponding longitude and latitude and height data are combined with the Beidou grid level to carry out corresponding coding of the Beidou grid coding set; s122, constructing a space geographic position As a parent node, a spatial geographic entity And performing space geographic association as child nodes.
9. The method for retrieving the entity space-time information and the topological relation based on the space-time knowledge graph of claim 2, wherein the grids in the Beidou satellite geographic map are subjected to space position resolution and hierarchical division, and a Beidou grid coding dataset is correspondingly divided into Beidou grid levels belonging to different scales up and down.
10. The method for retrieving spatial information and topological relation of entities based on spatiotemporal knowledge patterns according to claim 8, wherein said spatial geographic location and spatial geographic entity comprise spatial geographic information with spatial geographic types of points, lines and planes, the points represent spatial geographic location points, the lines represent linear sets of spatial geographic location points, and the planes represent planar sets of spatial geographic location points.

Description

Entity space-time information and topological relation retrieval method based on space-time knowledge graph Technical Field The invention relates to the field of space-time query and retrieval, in particular to a method for retrieving entity space-time information and topological relation based on a space-time knowledge graph. Background Along with the rapid development of artificial intelligence and big data technologies, knowledge maps are used as core technologies for integrating data and semantic association, and are widely applied to the fields of intelligent transportation, urban management, environmental monitoring and the like. Knowledge graph is an information extraction technology for extracting important entities and relations from mass information and constructing a graph net structure. Due to the fact that deep associated information is good at being mined and global semantics are understood, the output accuracy and reliability of retrieval enhancement (RETRIEVAL-Augmented Generation, RAG) are remarkably improved through the application of the knowledge graph, and the problems that knowledge of a large language model (Large Language Model, LLM) is outdated, illusion is generated and the like are solved. In recent years, a space-time knowledge graph (Spatio-Temporal Knowledge Graphs) technique has been developed. The research shows that constructing the space-time knowledge graph with space-time characteristic information can enhance the space-time semantic understanding capability and realize complex space-time reasoning (such as predicting the future spatial distribution of the entity and tracing time sequence association), thereby providing a direct decision basis for the scene service. However, the traditional space-time knowledge graph has two short plates in technical floor and business adaptation, namely, the first short plate has weak space-time reasoning capability and poor generalization. Traditional space-time reasoning is limited by simple time sequence association (for example, an event A occurs after an event B) or space proximity judgment (for example, a place C approaches to a place D), depth mining of complex space-time dependency relations of cross-scale (for example, urban planning and neighborhood level management and control), cross-scene (for example, daily monitoring and emergency scheduling) is lacked, and an existing space-time reasoning model is mainly based on fixed rule design or specific data set training, and is remarkably insufficient in model suitability and generalization capability when facing dynamically-changing actual business requirements (for example, resource scheduling scenes after sudden disasters). Secondly, the space identifiers are not uniform and the entity suitability is poor. On one hand, the traditional space-time knowledge graph does not introduce a unified space identification standard, lacks a unified space specification, has different space identification formats of different data sources, and causes the problems of easy coordinate conflict, repeated labeling, large data fusion difficulty and the like when the multi-source data are in butt joint, on the other hand, the traditional space-time knowledge graph is difficult to realize multi-scale adaptation for different types of geographic entities such as points (such as single monitoring equipment), lines (such as roads and rivers), faces (such as administrative areas and lakes) and the like, and particularly cannot accurately describe the space range and association relation of the geographic entities in irregular forms such as rivers and complex administrative areas and the like. Along with the rapid development of the knowledge graph technology, the graph retrieval technology is driven to develop to the accurate and intelligent direction, however, the prior knowledge graph retrieval technology still has obvious short plates under the space-time knowledge graph scene, and the method is specifically characterized in the following four aspects: Firstly, the complex semantic analysis capability is insufficient, the traditional knowledge graph retrieval method relies on accurate matching and single related word query, and the unstructured query requirements such as fuzzy time description (e.g. before and after flood season), fuzzy space description (e.g. around science and technology garden) and the like are difficult to analyze, so that the problems of insufficient mining of the core query intention of a user, low relevance of a retrieval result and the like are caused. Secondly, the multi-dimensional retrieval capability is insufficient, the traditional retrieval method generally ignores consideration on time-space dimension and cannot systematically process space-time information, and particularly, the retrieval accuracy is low and the bottleneck of low efficiency is particularly prominent in complex space-time scenes such as cross-scale space-time query (e.g. urban planning and neighborhood m