Analysis method, device, equipment, and storage medium for historical operation data

Abstract

The invention provides a historical operation data analysis method and device, equipment and a storage medium, and relates to the technical field of data processing, in particular to the technical field of intelligent transportation, automatic driving, large models and data analysis. According to the specific implementation scheme, at least one analysis anchor point is determined in a space-time knowledge graph representing the operation state of the unmanned carrying tool according to a to-be-analyzed tracing problem; starting from at least one analysis anchor point, tracing the incidence relation in the space-time knowledge graph so as to identify and construct a cause subgraph containing a potential cause-effect chain; converting the structured information of the incong sub-graph into an analysis prompt for a large model; and obtaining an analysis result of the tracing problem according to the reasoning of the analysis prompt by the large model. According to the technical scheme of the invention, the deep causal tracing of complex operation problems can be realized by combining a structured knowledge graph and a large language model with cognitive intelligence.

Description

(19) State Intellectual Property Office (12) Invention Patent Application (10) Application Publication Number (43) Application Publication Date (21) Application Number 202511351601.1 (22) Application Date 2025.09.19 (71) Applicant Apollo Intelligent Technology (Beijing) Co., Ltd. Address 105, 1st Floor, Building 1, No. 10, Shangdi 10th Street, Haidian District, Beijing 100085 (72) Inventors Yan Wenkai, Zhang Bo, Zhang Xiaojie (74) Patent Agency Beijing Yiguang Intellectual Property Agency Co., Ltd. 11596 Patent Attorney Jin Aijing, Wang Ying (51) Int.Cl. G06Q 10 / 047 (2023.01) G06Q 10 / 067 (2023.01) G06F 18 / 214 (2023.01) G06F 16 / 29 (2019.01) G06N 5 / 04 (2023.01) (54) Invention Title: Method, Apparatus, Device and Storage Medium for Analyzing Historical Operational Data (57) Abstract: This disclosure provides a method, apparatus, device and storage medium for analyzing historical operational data. This disclosure relates to the field of data processing technology, and particularly to the fields of intelligent transportation, autonomous driving, large models and data analysis technology. The specific implementation scheme is as follows: Based on the tracing problem to be analyzed, at least one analysis anchor point is determined in the spatiotemporal knowledge graph representing the operational status of unmanned vehicles; starting from at least one analysis anchor point, its correlation is traced in the spatiotemporal knowledge graph to identify and construct a causal subgraph containing potential causal chains; the structured information of the causal subgraph is transformed into analysis prompts for the large model; based on the reasoning of the large model on the analysis prompts, the analysis result of the tracing problem is obtained. According to the technical solution of this disclosure, a deep causal tracing of complex operational problems can be achieved by combining a structured knowledge graph and a large language model with cognitive intelligence. Claims 2 pages, Description 14 pages, Drawings 3 pages, CN 121235237 A 2025.12.30 CN 1 21 23 52 37 A 1. A method for analyzing historical operational data, comprising: determining at least one analysis anchor point in a spatiotemporal knowledge graph representing the operational status of a fleet, based on a tracing question to be analyzed; tracing the relationships between the at least one analysis anchor point in the spatiotemporal knowledge graph to identify and construct a causal subgraph containing potential causal chains; converting the structured information of the causal subgraph into analysis prompts for a large model; and obtaining the analysis result of the tracing question based on the reasoning of the large model on the analysis prompts. 2. The method according to claim 1, wherein determining at least one analysis anchor point in a spatiotemporal knowledge graph representing the operational status of a fleet, based on a tracing question to be analyzed, comprises: performing natural language understanding on the tracing question in text form to identify the tracing intent and extract key entities;3. The method according to claim 2, wherein locating at least one analysis anchor point in the spatiotemporal knowledge graph representing the fleet operation status based on the tracing intent and the key entity includes: when the tracing problem is an event-driven tracing problem, directly searching for the corresponding target event node in the spatiotemporal knowledge graph based on the event identifier in the key entity; and determining at least one entity node associated with the target event node as the analysis anchor point. 4. The method according to claim 2, wherein locating at least one analysis anchor point in the spatiotemporal knowledge graph based on the tracing intent and the key entity includes: when the tracing problem is a non-event-driven tracing problem, inferring and generating a set of operational event nodes related to the tracing problem in the spatiotemporal knowledge graph based on at least one of the entity identifier, time range, or indicator information of the key entity, as the at least one anchor point node. 5. The method according to claim 1, wherein tracing the relationships in the spatiotemporal knowledge graph starting from the at least one analysis anchor point to identify and construct a causal subgraph containing potential causal chains comprises: performing a temporal reverse traversal in the spatiotemporal knowledge graph starting from the at least one analysis anchor point to obtain multiple candidate cause nodes; scoring the causal contribution of the multiple candidate cause nodes; and filtering out core causal events based on the causal contribution scores to construct the causal subgraph. 6. The method according to claim 5, wherein scoring the causal contribution of the multiple candidate cause nodes comprises: calculating the causal contribution score based on at least one of the following: temporal proximity, spatial proximity, preset event association strength, and shared core business entities between the candidate cause nodes and the analysis anchor point. 7. The method according to claim 5, wherein the step of performing a temporal reverse traversal in the spatiotemporal knowledge graph starting from the at least one analysis anchor point to obtain multiple candidate cause nodes includes: setting a spatiotemporal constraint boundary based on the spatiotemporal attributes of the at least one analysis anchor point; performing a temporal reverse traversal in the spatiotemporal knowledge graph starting from any analysis anchor point; during the temporal reverse traversal, if the current node encountered exceeds the spatiotemporal constraint boundary, pruning the path where the current node is located until all paths have been traversed; obtaining multiple candidate cause nodes related to the tracing problem and containing potential causal chains based on the portion of the spatiotemporal knowledge graph retained after pruning; wherein the path includes nodes of the spatiotemporal knowledge graph and edges connected to the nodes.8. The method according to claim 1, wherein converting the structured information of the causal subgraph into analysis prompts for the large model comprises: converting the structured information of the causal subgraph into narrative text containing weighted evidence; embedding the narrative text and the tracing question into a preset analysis prompt template to obtain analysis prompts for the large model. 9. The method according to claim 8, wherein converting the structured information of the causal subgraph into narrative text containing weighted evidence comprises: assigning different strengths of evidence descriptions to the core causal events in the narrative text based on the causal contribution scores of the core causal events in the causal subgraph. 10. The method according to claim 1, wherein the spatiotemporal knowledge graph is constructed based on multi-source operational data reflecting the operational status of unmanned vehicles, obtained from at least multiple data sources including scheduling systems, vehicle-side equipment, and work order systems. 11. An apparatus for analyzing historical operational data, comprising: an extraction module for determining at least one analysis anchor point in a spatiotemporal knowledge graph characterizing the operational status of a fleet, based on a tracing problem to be analyzed; a construction module for tracing relationships in the spatiotemporal knowledge graph starting from the at least one analysis anchor point to identify and construct a causal subgraph containing potential causal chains; a transformation module for transforming the structured information of the causal subgraph into analysis prompts for a large model; and an inference module for obtaining the analysis result of the tracing problem based on the inference of the analysis prompts by the large model. 12. An electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1-10. 13. A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to perform the method according to any one of claims 1-10. 14. A computer program product, comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1-10. Claims 2 / 2 pages 3 CN 121235237 A Method, apparatus, device and storage medium for analyzing historical operating data Technical Field

[0001] This disclosure relates to the field of data processing technology, and more particularly to the fields of intelligent transportation, autonomous driving, large-scale models, and data analysis technology. Background Art

[0002] With the continuous advancement of autonomous driving technology, driverless vehicles, represented by driverless taxis, are gradually moving from the technology verification stage to large-scale commercial operation. In this process, the industry's focus has shifted from improving single-vehicle autonomous driving...The capability extends to how to ensure the safe, efficient, and economical operation of the entire fleet. Large-scale fleet operations generate massive amounts of multi-dimensional operational data, which provides a data foundation for achieving refined management and intelligent decision-making.

[0003] This disclosure provides a method, apparatus, device, and storage medium for analyzing historical operational data. According to one aspect of this disclosure, a method for analyzing historical operational data is provided, comprising:

[0004] determining at least one analysis anchor point in a spatiotemporal knowledge graph representing the fleet's operational status, based on the tracing problem to be analyzed;

[0005] tracing the relationships between the at least one analysis anchor point in the spatiotemporal knowledge graph to identify and construct a causal subgraph containing potential causal chains;

[0006] converting the structured information of the causal subgraph into analysis prompts for a large model;

[0007] obtaining the analysis results of the tracing problem based on the reasoning of the large model regarding the analysis prompts.

[0008] According to another aspect of the present disclosure, an analysis apparatus for historical operational data is provided, comprising:

[0009] an extraction module, configured to determine at least one analysis anchor point in a spatiotemporal knowledge graph characterizing the operational status of a fleet, based on a tracing problem to be analyzed;

[0010] a construction module, configured to trace the relationships between the at least one analysis anchor point in the spatiotemporal knowledge graph to identify and construct a causal subgraph containing potential causal chains;

[0011] a transformation module, configured to transform the structured information of the causal subgraph into analysis prompts for a large model;

[0012] a reasoning module, configured to obtain the analysis results of the tracing problem based on the reasoning of the analysis prompts by the large model.

[0013] According to another aspect of the present disclosure, an electronic device is provided, comprising:

[0014] at least one processor; and

[0015] a memory communicatively connected to the at least one processor; wherein,

[0016] the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform any of the methods in the embodiments of the present disclosure.

[0017] According to another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to cause the computer to perform any of the methods according to embodiments of the present disclosure.

[0018] According to another aspect of the present disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements any of the methods according to embodiments of the present disclosure.

[0019] According to the technical solutions of the present disclosure, a deep causal tracing of complex operational problems can be achieved by combining structured knowledge graphs and large language models with cognitive intelligence.

[0020] It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Brief Description of the Drawings

[0021] The drawings are provided to better understand the present invention and do not constitute a limitation of this disclosure. Wherein:

[0022] FIG1 is a flowchart illustrating a method for analyzing historical operating data provided in one embodiment of this disclosure;

[0023] FIG2 is a flowchart illustrating a method for analyzing historical operating data provided in another embodiment of this disclosure;

[0024] FIG3 is a flowchart illustrating a method for analyzing historical operating data provided in another embodiment of this disclosure;

[0025] FIG4 is a structural schematic diagram of a device for analyzing historical operating data provided in one embodiment of this disclosure;

[0026] FIG5 is a block diagram of an electronic device used to implement an embodiment of this disclosure. Detailed Description of Embodiments

[0027] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments of this disclosure to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0028] In related technologies, current operational data analysis mostly relies on expert systems or pre-set rule engines. Such systems can effectively identify known, clearly patterned single-point operational problems. However, when conducting in-depth operational analysis of unmanned vehicle fleets, especially when tracing the causal relationships of complex historical events, the existing technologies still have limitations. Many serious operational anomalies are often the result of the time-coupling effects of factors from multiple different fields and systems. For example, an increase in the user complaint rate in a region may be related to multiple dimensions of information, such as the vehicle hardware status in that region, recent adjustments to the dispatch strategy, and even the resource shortage of charging piles.

[0029] For such complex and unexpected operational problems, traditional expert systems or rule engines are difficult to effectively capture and diagnose using pre-set simple rules. Even if the data has been stored in a knowledge graph, the analysis process often relies on operators to perform manual, exploratory association queries and logical reasoning, making it difficult to automatically and deeply reveal the hidden causal chains behind the events.

[0030] To at least partially solve one or more of the above-mentioned problems and other potential problems, embodiments of this disclosure provide an analysis method for historical operational data based on knowledge graphs and large-scale models. Utilizing the technical solutions of these embodiments, a structured knowledge graph and a large-scale language model with cognitive intelligence can be combined to achieve deep causal tracing of complex operational problems.

[0031] This embodiment relates to a method for analyzing historical operational data, aiming to achieve deep causal tracing of complex operational problems by combining structured knowledge graphs and large language models with cognitive intelligence. Figure 1 is a flowchart of a method for analyzing historical operational data provided in an embodiment of this disclosure. As shown in Figure 1, the method includes at least the following steps:

[0032] S110, Based on the tracing problem to be analyzed, determine at least one analysis anchor point in the spatiotemporal knowledge graph representing the fleet's operational status.

[0033] In this embodiment, the fleet consists of multiple unmanned vehicles. Unmanned vehicles can be understood as autonomous vehicles that perform specific transportation tasks in a ground road network. Their specific forms include, but are not limited to: driverless taxis for passenger services and unmanned delivery vehicles for freight services. This embodiment is applicable to various unmanned vehicle operation scenarios.

[0034] For example, in an autonomous driving demonstration zone, there is a fleet of 150 pure electric driverless taxis. The operational goal of this fleet is to maximize the satisfaction of users' travel needs in the area while ensuring absolute safety. When a large number of driverless taxis are operated on a large scale, many operational problems will arise. For example, during morning and evening rush hours, a large number of orders will flow unidirectionally and centrally from residential areas to work areas (or in the opposite direction), resulting in a severe shortage of vehicles at the origin and a large backlog of vehicles at the destination. Another example is that due to untimely vehicle dispatching and unreasonable spatial distribution, the waiting time for users after placing an order is too long, affecting user experience and potentially leading to order cancellation.

[0035] In this step, it is first necessary to understand a traceability problem to be analyzed and find the starting point for analysis in a pre-built spatiotemporal knowledge graph.

[0036] A traceability problem can be understood as an analytical request proposed by operators or automated systems to explore the reasons behind a certain historical operational phenomenon. This problem can be in text form, such as "Analyze the reasons why order ORD-999 was complained about by users."

[0037] A spatiotemporal knowledge graph is a data model that uses a graph structure to model operational entities and the complex relationships between them. The "spatiotemporal" characteristic means that the entities and relationships in the graph have attributes with time and space dimensions. "Knowledge" is embodied in the fact that the graph contains not only raw data, but also deep causal or relational relationships extracted through rules and models. Conceptually, a knowledge graph consists of "entities" as nodes and "relationships" as edges. In this solution, the spatiotemporal knowledge graph can predefine entity types such as "vehicle," "order," "work order," and "spatiotemporal slice." Entity nodes refer to the nodes created in the graph database for each specific data record or operational event, based on its type. For example, when the system receives a vehicle...When data with vehicle ID "VN-123" is received, an entity node of type "vehicle" will be created, and "VN-123" will be used as its unique identifier. Operational data related to the vehicle can be used as attributes of the node or associated nodes connected to the node.

[0038] An analysis anchor point can be understood as one or more initial nodes located by the system in the spatiotemporal knowledge graph based on the tracing question, used to initiate subsequent causal chain tracing. The determination of the analysis anchor point is the bridge connecting the abstract question and the specific graph data. The specific determination method will be detailed in subsequent embodiments.

[0039] For example, suppose a tracing question to be analyzed is: "Analyze the reason why order ORD-999 was complained about by the user". After receiving the question, the system will first parse it and determine the node associated with "order ORD-999" in the spatiotemporal knowledge graph as the anchor point for this analysis.

[0040] S120. Starting from at least one analysis anchor point, trace its association in the spatiotemporal knowledge graph to identify and construct a causal subgraph containing potential causal chains.

[0041] In this step, after determining the analysis anchor point, the system will start from the anchor point and explore the graph to construct a local graph structure focused on causal analysis.

[0042] Tracing its relationships can be understood as a graph traversal or graph reasoning process, aiming to start from the anchor point and search along the edges in the graph for all other nodes that may have a causal relationship with it.

[0043] The causal subgraph can be understood as a subset extracted from the full knowledge graph. It contains the analysis anchor point and all potential causal nodes discovered by tracing relationships and the connection paths between them, thus forming a visualized network of potential causal chains.

[0044] For example, starting from the analysis anchor point "Order ORD-999", the system traces in the knowledge graph and may find that the order was executed by "Vehicle VN-456", and "Vehicle VN-456" experienced an "emergency braking" event 5 minutes before the order was executed. The system will construct a causal subgraph by combining the relationships between the "Order ORD-999" node, the "Vehicle VN-456" node, the "Emergency Braking" event node, and other nodes.

[0045] S130: Transform the structured information of the causal subgraph into analysis prompts for the large model.

[0046] In this step, the machine-readable graph structure data needs to be translated into a natural language format that the large language model (also commonly referred to as a large model or LLM) can understand.

[0047] The analysis prompt can be a well-designed text paragraph containing background information, context, and explicit instructions to guide the large model in high-quality reasoning.

[0048] For example, the causal subgraph information constructed in the previous step can be converted into text like this: "Background: Analyze the reason why order ORD-999 was complained about. Related event: The associated vehicle VN-456 experienced a sudden braking event 5 minutes before the service started." This text will become part of the analysis prompt.

[0049] S140. Based on the reasoning of the analysis prompt by the large model, obtain the analysis results of tracing the problem.

[0050] In this step, the analysis prompt generated in the previous step is sent to a pre-configured large language model through the API interface. The large model will perform logical reasoning based on the information in the prompt and return an analysis result in text form. The analysis result can be a comprehensive report that includes inferences about the root cause of the problem, logical explanations, confidence assessments, and suggestions for future improvement.

[0051] For example, after receiving the analysis prompt, the large model may reason and return: "The root cause of the complaint against order ORD-999 is most likely due to the sudden braking event that occurred on vehicle VN-456 during the pick-up process, resulting in a poor passenger experience."

[0052] According to the solution of the present disclosure embodiment, by determining the analysis anchor point, tracing the relationship to construct the causal subgraph, and then converting the subgraph information into analysis prompts for the large model to reason, a complete technical link from the original problem to the deep analysis result is established, which can combine the objectivity of structured graph data with the cognitive intelligence of the large language model, and solve the problem of complex and implicit causal relationship operation tracing that is difficult to handle by traditional methods.

[0053] In one possible implementation, step S110 determines at least one analysis anchor point in the spatiotemporal knowledge graph representing the fleet operation status according to the tracing problem to be analyzed, and may further include the following steps:

[0054] S111, perform natural language understanding on the tracing problem in text form to identify the tracing intent and extract key entities.

[0055] This step is a prerequisite for determining the analysis anchor point, aiming to understand the user's natural language questions. Natural language understanding can be implemented based on existing algorithms or models to parse text input. This process mainly includes two sub-tasks:

[0056] Tracing intent identification: Determine the type of user question, such as whether they want to trace a specific event or attribute a macro indicator.

[0057] Key entity extraction: Extract useful information fragments from the text, such as vehicle ID, order ID, time adverbs, location nouns, etc., and link them with standardized identifiers in the knowledge graph.

[0058] For example, for the question "Why did vehicle VN-123 go offline on Orchard Road yesterday?", the NLU module will identify:

[0059] Tracing intent: "Event-driven tracing".

[0060] Key entities: {Vehicle ID: "VN-123", Time: "yesterday", Location: "Orchard Road", Event: "offline"}.

[0061] S112. Based on the tracing intent and key entities, locate at least one analysis anchor point in the spatiotemporal knowledge graph.

[0062] In this step, the structured information parsed by NLU in the previous step is used to query the spatiotemporal knowledge graph to locate one or more nodes most suitable as the starting point of the analysis.

[0063] In the above example, based on the key entities {Vehicle ID:"VN-123", Time:"Yesterday", Event:"Offline"}, the knowledge graph can be queried for operational event nodes of type "offline" associated with vehicle VN-123 yesterday, and the node can be determined as the anchor point for this analysis.

[0064] According to the scheme of the embodiment of this disclosure, by introducing a natural language understanding module, the system can process unstructured text queries. The natural language queries of humans are accurately translated into machine-executable graph positioning instructions, which greatly improves the usability and intelligence level of the analysis system, enabling operators who are not familiar with database query languages ​​to perform in-depth data analysis.

[0065] In one possible implementation, step S112 locates at least one analysis anchor point in the spatiotemporal knowledge graph based on the tracing intent and key entities, and may further include the following steps:

[0066] In the case of an event-driven tracing problem, the corresponding target event node is directly searched in the spatiotemporal knowledge graph based on the event identifier in the key entity.

[0067] At least one entity node associated with the target event node is determined as the analysis anchor point.

[0068] In this embodiment of the present disclosure, an event-driven tracing problem can be understood as a tracing analysis centered around a clear operational event that has already occurred. When the NLU module determines that the problem belongs to this type, the system will execute a direct and efficient positioning strategy.

[0069] Specific example: For the question "Analyze the reason for customer complaint C-886", the key entity extracted by the NLU contains a unique event identifier C-886. The system will use this ID as an index to directly query and locate the target event node representing "customer complaint C-886" in the knowledge graph.

[0070] In some cases, tracing directly from the event node may not provide sufficient information. To obtain richer context, the system can use the core business entity directly associated with the event as the analysis anchor point. Entity nodes can represent physical or business objects, such as vehicle nodes, order nodes, etc.

[0071] In the above example, after finding the event node "Customer Complaint C-886", the system discovers that it is connected to the "Order ORD-999" node and the "Vehicle VN-456" node through a relationship edge. For comprehensive analysis, the system can jointly determine the two entity nodes "Order ORD-999" and "Vehicle VN-456" as the anchor points for this analysis.

[0072] According to the scheme of the embodiments of this disclosure, by designing a special anchor point positioning strategy for event-driven problems, the starting point of analysis can be found quickly and accurately. By using the core entities associated with the event as anchor points, it is ensured that the subsequent causal chain tracing can be carried out in a broader and more relevant context, thereby improving the comprehensiveness of the analysis.

[0073] In one possible implementation, step S112 locates at least one analysis anchor point in the spatiotemporal knowledge graph according to the tracing intent and key entities, and may further include the following steps:

[0074] When the tracing problem is a non-event-driven tracing problem, based on at least one of the entity identifier, time range, or indicator information of the key entity, a set of operational event nodes related to the tracing problem are inferred and generated in the spatiotemporal knowledge graph as at least one anchor point node.

[0075] In the embodiments of this disclosure, non-event-driven tracing problems can be understood as tracing analyses proposed for phenomena that do not have clear abnormal events but are aimed at a certain entity or macroscopic indicator. For such problems, the system needs to perform a "hypothesis generation" or "problem dimensionality reduction" process, that is, intelligently inferring which underlying events are most likely to have caused the macroscopic phenomena observed by the user.

[0076] Specific example: For the question "Why was the order volume of vehicle VN-789 very low yesterday?", there are no specific abnormal events. The system will query all related events of the vehicle yesterday in the knowledge graph based on the key entity {Vehicle ID: "VN-789", Time: "yesterday"}, and infer the most abnormal event, such as a "recharging in progress" event lasting 4 hours and several "order dispatch failure" events. The system will generate the anchor node set for this analysis by combining these inferred event nodes that are most likely to lead to the phenomenon of "low order volume". Specification 5 / 14 page 8 CN 121235237 A

[0077] According to the solution of the embodiment of this disclosure, by designing an intelligent inference and anchor point generation mechanism for non-event-driven problems, this method can handle more open and fuzzy analysis requests. This method can automatically reduce the dimensionality of a macroscopic, phenomenon-level problem and focus it on a set of specific, analyzable micro-events, greatly expanding the applicable scenarios and analytical capabilities of the method.

[0078] In one possible implementation, as shown in FIG2, step S120 starts from at least one analysis anchor point and traces its association in the spatiotemporal knowledge graph to identify and construct a causal subgraph containing potential causal chains. This may further include the following steps:

[0079] S210: Starting from at least one analysis anchor point, perform a temporal reverse traversal in the spatiotemporal knowledge graph to obtain multiple candidate cause nodes.

[0080] In this embodiment of the present disclosure, the temporal reverse traversal can be a graph search algorithm that starts from the analysis anchor point and traces along...Following the "originating direction" of the relationship edge, the system backtracks from the nearest to the furthest in time, searching for all nodes that occurred before the anchor point in time and have a path connection to the anchor point. Candidate cause nodes can be all nodes that meet the conditions found during this traversal, which constitute the initial set of potential causes.

[0081] For example, starting from the anchor point "order ORD-999 was complained about", the system traverses backward and may find multiple candidate cause nodes such as "vehicle VN-456 executed this order", "vehicle VN-456 braked suddenly at T-5 minutes", and "vehicle VN-456 replanned its route at T-10 minutes".

[0082] S220, score the causal contribution of multiple candidate cause nodes.

[0083] In this embodiment of the disclosure, in order to select the most important cause from many candidate causes, it is necessary to quantitatively evaluate each candidate cause node. The causal contribution score can be a comprehensive score used to measure the probability that there is a real causal relationship between a candidate cause node and the analysis anchor point.

[0084] In the above example, the "sudden braking" node and the "path replanning" node found in the previous step will be scored respectively. It is possible that the causal contribution score of "sudden braking" is 0.92, while the score of "path replanning" is 0.75.

[0085] S230. Select core causal events based on the causal contribution scores to construct a causal subgraph.

[0086] In this embodiment of the present disclosure, a scoring threshold can be set to filter out all candidate cause nodes with scores higher than the threshold. Core causal events are these high-probability cause nodes that have passed the scoring screening. The final causal subgraph is composed of the analysis anchor points, these core causal events, and the relationships between them.

[0087] If the scoring threshold is 0.7, then the above two nodes, "sudden braking" and "path replanning", will be selected as core causal events and included in the final causal subgraph.

[0088] According to the scheme of this disclosure embodiment, through the three-step process of "traversal-scoring-screening", the most noteworthy core reasons can be intelligently and quantitatively identified from massive amounts of related information. This method replaces manual experience judgment with a data-driven scoring model, greatly improving the objectivity and accuracy of cause location.

[0089] In one possible implementation, step S220 scores the causal contribution of multiple candidate cause nodes, which may further include the following steps:

[0090] Calculate the causal contribution score based on at least one of the following: temporal proximity, spatial proximity, preset event correlation strength, and shared core business entities between the candidate cause node and the analysis anchor point.

[0091] In this disclosure embodiment, the causal contribution score can be a comprehensive score calculated by weighting multiple dimensions, which can be any one, two, three, four, or even any number of more dimensions. These dimensions can be:

[0092] Temporal proximity: The smaller the time difference between the candidate cause and the anchor point, the higher the score. (Instruction manual 6 / 14 page 9 CN 121235237 A)

[0093] Spatial proximity: The closer the two are geographically, the higher the score.

[0094] Preset event association strength: Based on offline mining of massive historical data, the association probability between different event types is pre-calculated. For example, historical data shows that the probability of a "complaint" occurring after "sudden braking" is 30%, so this 30% can be used as part of the association strength score.

[0095] Shared core business entities: If the candidate cause and the anchor point are associated with the same vehicle or the same order, a higher score can be obtained.

[0096] Taking the above example, for the candidate cause "sudden braking" event, it is calculated that its time difference with the "complaint" anchor point is only 5 minutes (high time score), the location is consistent (high spatial score), the historical association is strong, and it shares the same vehicle and order (high entity sharing score), ultimately resulting in a high comprehensive score of 0.92.

[0097] According to the scheme of the present disclosure embodiment, through a multi-dimensional scoring model, the possibility of an event becoming a cause can be comprehensively and holistically evaluated from multiple perspectives such as time, space, historical statistics, and business logic. This comprehensive quantitative evaluation method is more reliable and robust than single-dimensional judgment.

[0098] In one possible implementation, as shown in FIG3, step S210 starts from at least one analysis anchor point and performs a temporal reverse traversal in the spatiotemporal knowledge graph to obtain multiple candidate cause nodes, and may further include the following steps:

[0099] S211, setting spatiotemporal constraint boundaries according to the spatiotemporal attributes of at least one analysis anchor point.

[0100] In the present disclosure embodiment, before the traversal begins, in order to improve efficiency, a reasonable search range is first defined by the spatiotemporal constraint boundaries. The spatiotemporal constraint boundaries can be a search range jointly defined by time and space, such as: a "30-minute" time window ending at the time of occurrence of the analysis anchor point, and a "5-kilometer" radius range centered on the location of the anchor point.

[0101] S212. Starting from any analysis anchor point, perform a temporal reverse traversal in the spatiotemporal knowledge graph.

[0102] S213. During the temporal reverse traversal, if the current node encountered exceeds the spatiotemporal constraint boundary, prune the path of the current node until all paths have been traversed.

[0103] In this embodiment of the present disclosure: pruning means that during the traversal, when the algorithm finds that the occurrence time or location of a candidate cause node has exceeded the set boundary, it will immediately stop tracing back along that node.

[0104] For example, starting from the "complaint" anchor point (occurring at 3:00 PM, location A), when traversing to a node where the occurrence time or location exceeds the set boundary, prune the path of the current node until all paths have been traversed.During the "vehicle maintenance" event at 10:00 AM, since its time has exceeded the boundary of "the first 30 minutes", any events that occurred before "vehicle maintenance" will no longer be explored, thus pruning this invalid search path.

[0105] S214. Based on the partial spatiotemporal knowledge graph retained after pruning, multiple candidate cause nodes related to the tracing problem and containing potential causal chains are obtained. Among them, the path includes nodes of the spatiotemporal knowledge graph and edges connected to the nodes.

[0106] According to the scheme of the embodiment of this disclosure, by presetting the spatiotemporal constraint boundary and performing dynamic pruning in the traversal, the computing resources can be concentrated in the most relevant time and space range. This method greatly reduces unnecessary graph search overhead and significantly improves the analysis efficiency of causal tracing under large-scale knowledge graphs.

[0107] In one possible implementation, step S130 transforms the structured information of the subgraph into analysis prompts for the large model, and may further include the following steps:

[0108] S131. Transform the structured information of the causal subgraph into narrative text containing weighted evidence.

[0109] In this embodiment of the disclosure, the narrative text can be a natural language text that simulates a human analyst writing a report and has logic and readability. The text organizes the nodes and edges in the causal subgraph according to a certain logic (such as chronological order, importance order). Weighted evidence refers to the text not only stating the facts, but also clearly indicating the strength of the correlation between the facts (i.e., the core causal event) and the problem. This strength can be derived from the causal contribution score. Specification 7 / 14 pages 10 CN 121235237 A

[0110] S132. Embed the narrative text and the tracing problem into a preset analysis prompt template to obtain analysis prompts for the large model.

[0111] In this embodiment of the disclosure, the analysis prompt template can be a pre-designed text framework that includes role-playing, task instructions, and output format requirements. Filling the "background information" part of the template with the narrative text generated in the previous step can constitute a complete and high-quality analysis prompt.

[0112] For example, an analysis prompt template could be: "#Role: You are a senior operations analysis expert. #Background Information: {fill in the narrative text generated in S131 here}. #Task: Based on the background information, please provide the root cause and improvement suggestions."

[0113] According to the scheme of this disclosure embodiment, by converting the structured causal subgraph into weighted narrative text and combining it with the analysis prompt template, a high-quality way to interact with large models is created. This method provides highly focused objective evidence that has been preprocessed by the algorithm to the large model in a form best suited for the large model's understanding and reasoning, which is a key step in ensuring the accuracy and depth of subsequent analysis results.

[0114] In one possible implementation, step S131 transforms the structured information of the causal subgraph into narrative text containing weighted evidence, and may further include the following steps:

[0115] S310, assigning different evidence strength descriptions to the core causal events in the narrative text based on the causal contribution scores of the core causal events in the causal subgraph.

[0116] In this embodiment of the disclosure, this step is a specific implementation of "weighted evidence". A mapping rule from score to description can be set.

[0117] Score > 0.9 -> Evidence strength description can be: "Highly relevant core evidence".

[0118] 0.7 < Score <= 0.9 -> Description can be: "Significantly relevant possible cause".

[0119] 0.5 < Score <= 0.7 -> Description can be: "Auxiliary factor worthy of attention".

[0120] Specific example: In embodiment 5, the score of the "sudden braking" event is 0.92, and the score of the "path replanning" event is 0.75. In the generated narrative text, they are described as follows:

[0121] "[Highly relevant core evidence]: About 1 minute before the complaint occurred, the vehicle involved experienced a 'sudden braking' event (overall contribution 0.92)."

[0122] "[Significantly relevant possible cause]: About 5 minutes before the complaint occurred, the vehicle performed a 'path replanning' event (overall contribution 0.75)."

[0123] According to the scheme of the embodiments of this disclosure, by assigning a description of the strength of evidence with quantitative basis to different core causal events, the transformed narrative text not only contains facts, but also conveys the importance level of these facts. This approach greatly reduces the reasoning burden of large models, enabling them to quickly grasp the key to the problem and conduct logical analysis around the core evidence, thereby improving the accuracy and logic of the final analysis results.

[0124] In one possible implementation, the spatiotemporal knowledge graph is constructed based on multi-source operational data reflecting the operational status of unmanned vehicles, obtained from multiple data sources, including at least scheduling systems, vehicle-side equipment, and work order systems.

[0125] In this embodiment of the disclosure, data is first collected from various key systems in the operational ecosystem. Multi-source operational data refers to a collection of data generated in the daily operation of unmanned vehicles, which has diverse sources and different formats. These data sources include at least:

[0126] Scheduling system: providing order information, trip information (including the start and end locations of the trip), vehicle-order matching records, dispatch strategy logs, pricing information, etc., usually in the form of structured database records.

[0127] Vehicle-side equipment: mainly refers to the vehicle's telematics processor and autonomous driving computing unit (MPU), which will provide real-time data.The system reports vehicle location (GPS), speed, acceleration, battery charging status, vehicle hardware status, fault diagnostic codes, etc., and may also include vehicle trajectories formed by the geographical location information of the vehicle at multiple consecutive moments, which are usually high-frequency real-time data streams.

[0128] Work order system: records all offline operation and maintenance activities, such as charging work orders, battery swapping work orders, maintenance work orders, cleaning work orders, etc., and includes information such as activity type, creation time, execution location, and completion status.

[0129] The unmanned vehicle operation status includes two levels: micro-level individual status data and macro-level global situation data:

[0130] 1. Micro-level individual status data, mainly refers to the specific and independent business link status of each unmanned vehicle. This data mainly comes from the vehicle-side equipment.

[0131] 2. Macro-level global situation data, mainly refers to the overall data reflecting the current supply and demand, resource distribution and efficiency from the perspective of the entire fleet or a specific operating area. This data mainly comes from the real-time statistics and summaries of the scheduling system and the work order system.

[0132] The "spatiotemporal" characteristic in the spatiotemporal knowledge graph represents that the entities and relationships in the graph model have attributes of time and space dimensions, enabling them to describe the evolution of the operational state in spatiotemporal space. "Knowledge" is reflected in the fact that the graph not only contains raw data, but also deep causal or correlational relationships extracted through rules and models. This knowledge graph is conceptually composed of "entities" as nodes and "relationships" as edges.

[0133] It can be understood that the spatiotemporal knowledge graph is obtained by deeply integrating and associating multi-source data from all corners of the unmanned vehicle operation ecosystem through a series of steps such as data access, entity recognition, and relationship construction.

[0134] According to the scheme of the embodiment of this disclosure, by constructing a spatiotemporal knowledge graph based on multi-source operational data, it is ensured that all analysis and prediction are based on comprehensive, real-time, and accurate data, guaranteeing the reliability and practical value of the final output results.

[0135] Based on the aforementioned implementation steps, two complete examples corresponding to "event-driven" and "non-event-driven" tracing problems are given to illustrate the actual application process of this scheme in detail.

[0136] Example 1:

[0137] First, let's describe the business scenario of this example. Yesterday at noon peak, at 12:30, the system automatically generated and alerted a "severe supply-demand imbalance" event. This event occurred within a geographic grid in the Central Business District (CBD). Operations analysts need to investigate why such a severe capacity shortage occurred at this critical time and location.

[0138] The first step in the analysis process is to obtain and analyze the traceability question. After the operations personnel or analyst selects the alert event in the system, it will be transformed into a specific traceability question: "Analyze the event that occurred yesterday at 12:30 in the CBD grid (w21z7)..."The root cause of the 'severe supply and demand imbalance' event (ID: IMB-554). The natural language understanding module will parse the input, identify its tracing intent as "event-driven tracing", and extract key entities, including the event ID "IMB-554", the time "yesterday 12:30", and the location "w21z7".

[0139] Next, we proceed to the step of determining the analysis anchor point. Since the problem is event-driven, we can directly locate the operational event node of the "severe supply and demand imbalance" in the knowledge graph based on the event ID "IMB-554". The system finds that the event node is closely related to the "spatiotemporal slice" node "w21z7@yesterday 12:30" representing the spatiotemporal range. In order to analyze the cause of the spatiotemporal slice imbalance, this spatiotemporal slice node is determined as the anchor point for this analysis.

[0140] After the anchor points were determined, the causal subgraph was constructed. Starting from the anchor point “w21z7@yesterday12:30”, a reverse traversal of the time sequence was performed, focusing on finding events that negatively impacted the capacity supply of the area before the imbalance occurred, i.e., between 12:00 and 12:30. The system discovered and scored several candidate causes. One candidate cause was the work order event of “fleet concentrated charging”, which showed that 5 vehicles collectively swapped batteries at the edge of the CBD at 12:15, with a causal contribution score of 0.90. Another candidate cause was the “traffic congestion” event from third-party data on page 9 / 14 of the manual, CN 121235237 A, which showed that a main road leading into the CBD was severely congested at 12:00, with a causal contribution score of 0.85. Yet another candidate cause was the “schedule strategy change” event that occurred at 11:00 am, which slightly reduced the incentive weight for idle vehicles to actively cruise, with a causal contribution score of 0.60.

[0141] After setting the scoring threshold to 0.5, the above three reasons were all selected as core causal events. Based on this, the system constructed a causal subgraph containing nodes such as "supply and demand imbalance", "centralized energy replenishment", "traffic congestion", and "strategy change".

[0142] Subsequently, the system converted the constructed causal subgraph into analysis prompts for the large model. The conversion process is to generate a narrative text from the structured information of the graph, for example: "Background: Analysis of the reasons for the severe supply and demand imbalance that occurred in the CBD grid (w21z7) at 12:30 yesterday. Highly relevant core evidence: 15 minutes before the imbalance occurred, 5 vehicles near the grid collectively entered the 'battery swapping' state, resulting in a sudden reduction in supply (overall contribution 0.90). Significantly relevant possible cause: 30 minutes before the imbalance occurred, a 'severe congestion' event occurred on a main road entering the grid, hindering the inflow of external transportation capacity (overall contribution 0.85). Auxiliary factors worth noting: In the morning of that day, the incentive weight of the 'active cruise of idle vehicles' scheduling strategy was reduced (overall contribution 0.60)." "

[0143] "This narrative text is then embedded into a preset analysis prompt template to form the final analysis prompt.

[0144] The final step is to obtain the analysis results. After receiving the analysis prompt, the large language model performs reasoning and outputs the root cause inference and operational improvement suggestions. The root cause inference is: This supply and demand imbalance is a typical operational risk caused by multiple concurrent factors. The direct cause is the superposition of a concentrated fleet replenishment behavior and a traffic congestion on a key road before the noon peak, which simultaneously reduces both the existing supply and the incremental supply. The earlier adjustment of the scheduling strategy reduced the system's capacity reserves, making the system less resilient to this concurrent impact. The operational improvement suggestions include: it is recommended to coordinate with the replenishment plans of large customer fleets to guide them to replenish their fleets outside of the noon and evening peak hours; at the same time, the scheduling system should more sensitively combine real-time traffic data and automatically increase the scheduling incentives for vehicles on surrounding uncongested roads when the main roads entering the hot area are congested.

[0145] Example 2: Non-Event-Driven Tracing – Attribution Analysis of “Regional Service Quality Decline”

[0146] This example aims to answer a more open, macro-level tracing question arising from operational report data.

[0147] First, the business scenario of this example is described. The operations manager discovered in the weekly report that the “average passenger waiting time” in area A (Jurong East) was as high as 12 minutes last Saturday, far exceeding the average level of 7 minutes, and the reasons need to be investigated.

[0148] The first step in the analysis process is to obtain and parse the tracing question. The operations manager inputs the question into the system: “Why was the average passenger waiting time in area A higher than usual last Saturday?” The system’s natural language understanding module parses the input, identifies its tracing intent as “non-event-driven tracing,” and extracts key entities, including location “A,” time “last Saturday,” and the indicator “average passenger waiting time” being higher than usual.

[0149] Next, the step of determining the analysis anchor point is taken. Since the question is non-event-driven, the system will perform an intelligent inference. The system first queries the knowledge graph ontology to find key driving factors affecting the "average passenger waiting time," such as "regional supply-demand ratio," "average pick-up distance," and "order dispatch failure rate." By comparing the data of these factors in region A for "last Saturday" and "the previous Saturday," the system finds that the "average pick-up distance" was 80% higher last Saturday than the previous Saturday, making it the most significant factor. At this point, the problem is automatically reduced in dimensionality and focused on "analyzing the reasons for the excessively long average pick-up distance in region A last Saturday." The system further queries all order events in the region that day with a "pick-up distance" greater than 3 kilometers, and intelligently infers and generates these specific "long-distance order dispatch" events as the set of analysis anchor points for this analysis.

[0150] After the anchor points are determined, the system begins to construct a causal subgraph. The system starts from these "long-distance order dispatch" event anchor points...The system performs a reverse traversal of the time sequence and scores the candidate causes found. One candidate cause is the "weather" event node, which shows that there was a "heavy rain" record in area A from 2 pm to 5 pm that day, with a causal contribution score of 0.92. Another candidate cause is that, according to the analysis of the trajectories of the order-taking vehicles on page 10 / 14 of the manual, CN 121235237 A, a large number of order-taking vehicles originated from the neighboring area B, which directly explains the long pick-up distance, with a causal contribution score of 0.98.

[0151] Based on these core causes, the system constructs a complete causal subgraph from "heavy rain" to "decreased local vehicle supply in A" to "cross-regional dispatch from B", ultimately leading to "excessively long average pick-up distance" and "high average waiting time".

[0152] Subsequently, the system converts the causal subgraph into analysis prompts. The generated narrative text is: "Background: Analysis of the reasons for the high average waiting time in area A last Saturday. The core phenomenon is the 'excessive average pick-up distance' in this area. Highly relevant core evidence: Analysis shows that a large number of orders were dispatched from idle vehicles in the neighboring area B over long distances (overall contribution 0.98). Significantly relevant possible reasons: This phenomenon is highly correlated with the 'heavy rain' weather event that occurred in A from 2 pm to 5 pm that day. Historical data shows that heavy rain is the main reason for the temporary shortage of local transportation capacity, which in turn triggers large-scale cross-regional dispatch (overall contribution 0.92)."

[0153] This narrative text was then embedded in the analysis prompt template.

[0154] Finally, the analysis results were obtained. After receiving the prompt, the large language model performed inference and output the root cause inference and operation improvement suggestions. The root cause inference is: The direct cause of the high waiting time is the excessive average pick-up distance, while the root cause is that during the heavy rain, area A experienced a local and temporary shortage of transportation capacity, which forced the dispatch system to call up vehicles from neighboring areas on a large scale to meet the demand. The operation improvement suggestions are not elaborated here.

[0155] It should be noted that the two examples above are for illustrating the basic process of this method in a clear way. In actual large-scale operations, the situation is much more complex. The causal subgraph of an operational anomaly event (such as "regional supply and demand imbalance") often does not contain only two or three clearly pointing candidate causes, but may contain dozens or even hundreds of spatiotemporally related event nodes with different scores, forming a complex network. The contribution score obtained by the algorithm essentially measures strong correlation, not true causation. In a complex subgraph, there may be multiple events with very high scores. For example, a "vehicle malfunction" and a "traffic congestion" may both have high scores, but whether the congestion caused the vehicle to idle for too long and thus triggered a malfunction alarm, or whether the vehicle malfunction caused the traffic congestion, the algorithm score itself cannot make such a deep logical and common-sense judgment.

[0156] Real operational problems are usually not caused by a single reason, but are the result of the interaction and amplification of multiple factors. For example, a slight "schedule strategy adjustment" (score 0.6), a "light rain" (score 0.5), and a few vehicles "charging normally" (score 0.4) are not the main causes on their own. But their combination—the strategy adjustment leading to a slight uneven distribution of vehicles, the light rain slowing down the overall vehicle speed, and a few vehicles going to charge at the same time—leads to a serious supply and demand imbalance. Traditional scoring models have difficulty capturing this non-linear correlation and combination.

[0157] Moreover, algorithmic scoring cannot understand the deep semantics and business implications of events. It cannot truly understand the semantics of "detours" in different scenarios. Is it a "reasonable detour" to avoid congestion, or an "abnormal detour" due to a navigation system error? Or is it a "compliant detour" to meet passengers' temporary changes of destination? These all require judgment based on a broader context and common sense, which is beyond the scope of scoring models.

[0158] Therefore, in this complex fog, large language models play a huge cognitive and reasoning role that algorithms cannot replace. It is responsible for reasoning from "relevance" to "causality". The analysis prompts it receives are not just a list of events, but a narrative text that already contains weighted evidence. The big model can examine this "list of evidence" like an experienced operations expert, and combine its built-in world knowledge and logical reasoning ability to construct the most logical causal narrative. It can determine the cause and effect in a specific scenario, thus completing the causal judgment that the scoring model cannot complete.

[0159] It can be understood that the reverse traversal and contribution scoring of the knowledge graph are responsible for quickly and accurately collecting and organizing all relevant, weighted suspicious clues from massive amounts of data. The big language model is responsible for the final and decisive logical reasoning and judgment of these complex clues.

[0160] Figure 4 is a structural schematic diagram of a historical operation data analysis device 400 provided according to an embodiment of the present disclosure. As shown in Figure 4, the device includes at least:

[0161] an extraction module 401, configured to determine at least one analysis anchor point in a spatiotemporal knowledge graph representing the fleet operation status based on the tracing problem to be analyzed;

[0162] a construction module 402, configured to trace the relationships between the at least one analysis anchor point in the spatiotemporal knowledge graph to identify and construct a causal subgraph containing potential causal chains;

[0163] a transformation module 403, configured to transform the structured information of the causal subgraph into analysis prompts for the large model;

[0164] a reasoning module 404, configured to obtain the analysis results of the tracing problem based on the reasoning of the analysis prompts by the large model.

[0165] In one possible implementation, the extraction module 401 is used for:

[0166] Natural language understanding is performed on the text-based tracing problem to identify the tracing intent and extract key entities;

[0167] Based on the tracing intent and key entities, at least one analysis anchor point is located in the spatiotemporal knowledge graph.

[0168] In one possible implementation, the extraction module 401 is used to:

[0169] When the tracing problem is an event-driven tracing problem, the corresponding target event node is directly searched in the spatiotemporal knowledge graph based on the event identifier in the key entity;

[0170] At least one entity node associated with the target event node is determined as the analysis anchor point.

[0171] In one possible implementation, the extraction module 401 is used to:

[0172] When the tracing problem is a non-event-driven tracing problem, a set of operational event nodes related to the tracing problem is inferred and generated in the spatiotemporal knowledge graph based on at least one of the entity identifier, time range, or indicator information of the key entity, as at least one anchor point node.

[0173] In one possible implementation, the construction module 402 is used to:

[0174] Starting from at least one analysis anchor point, perform a temporal reverse traversal in the spatiotemporal knowledge graph to obtain multiple candidate cause nodes;

[0175] Score the causal contribution of the multiple candidate cause nodes;

[0176] Filter out core causal events based on the causal contribution scores to construct a subgraph.

[0177] In one possible implementation, the construction module 402 is used to:

[0178] Calculate the causal contribution score based on at least one of the following: temporal proximity, spatial proximity, preset event association strength, and shared core business entities between the candidate cause nodes and the analysis anchor points.

[0179] In one possible implementation, the construction module 402 is used to:

[0180] set a spatiotemporal constraint boundary based on the spatiotemporal attributes of at least one analysis anchor point;

[0181] starting from any analysis anchor point, perform a temporal reverse traversal in the spatiotemporal knowledge graph;

[0182] during the temporal reverse traversal, if the current node being traversed exceeds the spatiotemporal constraint boundary, prune the path where the current node is located until all paths have been traversed;

[0183] based on the portion of the spatiotemporal knowledge graph retained after pruning, obtain multiple candidate cause nodes related to the tracing problem and containing potential causal chains; wherein, the path includes nodes of the spatiotemporal knowledge graph and edges connected to the nodes.

[0184] In one possible implementation, the transformation module 403 is used to:

[0185] transform the structured information of the causal subgraph into narrative text containing weighted evidence;

[0186] embed the narrative text and the tracing problem into a preset analysis prompt template to obtain analysis prompts for the large model. Specification page 12 / 14 15 CN 121235237 A

[0187] In one possible implementation, the conversion module 403 is used for:

[0188] Based on the causal contribution score of the core causal event in the causal subgraph, different strengths of evidence are assigned to the core causal events in the narrative text.

[0189] The specific functions and examples of each module and submodule of the apparatus in the embodiments of this disclosure can be found in the relevant descriptions of the corresponding steps in the above method embodiments, and will not be repeated here.

[0190] In the technical solutions of this disclosure, the acquisition, storage and application of user personal information involved all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0191] According to the embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0192] FIG5 shows a schematic block diagram of an example electronic device 500 that can be used to implement the embodiments of this disclosure. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0193] As shown in FIG5, device 500 includes a computing unit 501, which can perform various appropriate actions and processes according to a computer program stored in read-only memory (ROM) 502 or a computer program loaded from storage unit 508 into random access memory (RAM) 503. Various programs and data required for the operation of device 500 may also be stored in RAM 503. The computing unit 501, ROM 502, and RAM 503 are interconnected via bus 504. Input / output (I / O) interface 505 is also connected to bus 504.

[0194] Multiple components in device 500 are connected to I / O interface 505, including: input unit 506, such as keyboard, mouse, etc.; output unit 507, such as various types of displays, speakers, etc.; storage unit 508, such as disk, optical disk, etc.; and communication unit 509, such as network card, modem, wireless transceiver, etc. Communication unit 509 allows device 500 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0195] Computing unit 501 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of computing unit 501 include, but are not limited to, central processing unit (CPU), graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processor (DSP), and others.Any suitable processor, controller, microcontroller, etc. The computing unit 501 executes the various methods and processes described above, such as methods for analyzing historical operating data. For example, in some embodiments, the methods for analyzing historical operating data may be implemented as computer software programs tangibly contained in a machine-readable medium, such as storage unit 508. In some embodiments, part or all of the computer program may be loaded and / or installed on device 500 via ROM 502 and / or communication unit 509. When the computer program is loaded into RAM 503 and executed by computing unit 501, one or more steps of the methods for analyzing historical operating data described above may be performed. Alternatively, in other embodiments, computing unit 501 may be configured to execute methods for analyzing historical operating data by any other suitable means (e.g., by means of firmware).

[0196] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SOCs), payload programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transferring data and instructions to the storage system, the at least one input device, and the at least one output device.

[0197] Program code for implementing the methods of this disclosure can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a standalone software package, or entirely on a remote machine or server.

[0198] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include, but is not limited to, electrical...Machine-readable storage media may be a magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination thereof. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0199] To provide interaction with a user, the systems and techniques described herein may be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with a user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0200] The systems and techniques described herein can be implemented in computing systems that include back-end components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include front-end components (e.g., a user computer with a graphical user interface or a web browser through which the user can interact with embodiments of the systems and techniques described herein), or computing systems that include any combination of such back-end components, middleware components, or front-end components. Components of the system can be interconnected via digital data communication (e.g., a communication network) of any form or medium. Examples of communication networks include: local area networks (LANs), wide area networks (WANs), and the Internet.

[0201] Computer systems can include clients and servers. Clients and servers are generally geographically distant from each other and typically interact via communication networks. A client-server relationship is created by computer programs running on corresponding computers and having a client-server relationship with each other. The server can be a cloud server, a server for a distributed system, or a server incorporating blockchain technology.

[0202] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure are achieved, and no limitation is imposed herein.

[0203] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the principles of this disclosure should be included within the scope of protection of this disclosure. Instruction sheet 14 / 14 page 17 CN 121235237 A Figure 1 Figure 2 Instruction sheet with drawing 1 / 3 page 18 CN 121235237 A Figure 3 Figure 4 Instruction sheet with drawing 2 / 3 page 19 CN 121235237 A Figure 5 Instruction sheet with drawing 3 / 3 page 20 CN 121235237 A Abstract Abdominal ultrasound examination method, system and device Abstract This disclosure provides an analysis method, device, equipment, and storage medium for historical operation data. It pertains to the field of data processing technology, specifically intelligent transportation, autonomous driving, large models, and data analysis technology. The specific implementation plan is as follows: Based on the retrospective problem to be analyzed, at least one analysis anchor point is determined in the spatiotemporal knowledge graph representing the operational status of unmanned vehicles.and construct a causal subgraph containing potential causal chains. The structured information of the causal subgraph is converted into analysis prompts for the large model. Based on the inference of the large model on the analysis prompts, the analysis result of the retrospective problem is obtained. According to the technical solution of this disclosure, it is possible to combine the structured knowledge graph and a large language model with cognitive intelligence to achieve deep causal tracing of complex operational issues.

Claims

1. An analysis method of historical operation data, comprising: determining at least one analysis anchor point in a time-space knowledge graph representing a state of a fleet operation according to a trace-back problem to be analyzed; tracing back associated relationships in the time-space knowledge graph from the at least one analysis anchor point to identify and construct a causal sub-graph containing a potential causal chain; transforming structured information of the causal sub-graph into an analysis prompt for a large model; obtaining an analysis result of the trace-back problem according to reasoning of the large model on the analysis prompt.

2. The method of claim 1, wherein, The determining at least one analysis anchor point in a time-space knowledge graph representing a state of a fleet operation according to a trace-back problem to be analyzed comprises: performing natural language understanding on the trace-back problem in text form to identify a trace-back intention and extract key entities; locating at least one analysis anchor point in a time-space knowledge graph representing a state of a fleet operation according to the trace-back intention and the key entities.

3. The method of claim 2, wherein, The locating at least one analysis anchor point in a time-space knowledge graph representing a state of a fleet operation according to the trace-back intention and the key entities comprises: in a case where the trace-back problem is an event-driven trace-back problem, directly searching for a target event node in the time-space knowledge graph according to an event identifier in the key entities; determining at least one entity node associated with the target event node as an analysis anchor point.

4. The method of claim 2, wherein, The locating at least one analysis anchor point in a time-space knowledge graph representing a state of a fleet operation according to the trace-back intention and the key entities comprises: in a case where the trace-back problem is a non-event-driven trace-back problem, inferring and generating a group of operation event nodes related to the trace-back problem in the time-space knowledge graph according to at least one of an entity identifier, a time range or index information of the key entities as the at least one anchor node.

5. The method of claim 1, wherein, The tracing back associated relationships in the time-space knowledge graph from the at least one analysis anchor point to identify and construct a causal sub-graph containing a potential causal chain comprises: performing time sequence reverse traversal in the time-space knowledge graph from the at least one analysis anchor point to obtain a plurality of candidate cause nodes; performing causal contribution degree scoring on the plurality of candidate cause nodes; filtering out core causal events according to the causal contribution degree scoring to construct the causal sub-graph.

6. The method of claim 5, wherein, The performing causal contribution degree scoring on the plurality of candidate cause nodes comprises: calculating the causal contribution degree scoring according to at least one of time proximity, space proximity, a preset event association strength and a shared core business entity between the candidate cause nodes and the analysis anchor point.

7. The method of claim 5, wherein, The performing time sequence reverse traversal in the time-space knowledge graph from the at least one analysis anchor point to obtain a plurality of candidate cause nodes comprises: setting a time-space constraint boundary according to time-space attributes of the at least one analysis anchor point; performing time sequence reverse traversal in the time-space knowledge graph from any analysis anchor point; in a case where a current node traversed exceeds the time-space constraint boundary in the time sequence reverse traversal process, pruning a path where the current node is located until all paths are traversed; According to the part of the spatio-temporal knowledge graph retained after pruning, a plurality of candidate cause nodes related to the traceability problem and containing potential causal chains are obtained; wherein, the path includes nodes of the spatio-temporal knowledge graph and edges connected to the nodes.

8. The method of claim 1, wherein, The structured information of the causal subgraph is converted into an analysis prompt for the large model, including: The structured information of the causal subgraph is converted into a narrative text containing weighted evidence. The narrative text and the traceability problem are embedded into a preset analysis prompt template to obtain an analysis prompt for the large model.

9. The method of claim 8, wherein, The structured information of the causal subgraph is converted into a narrative text containing weighted evidence, including: According to the causal contribution score of the core causal event in the causal subgraph, the core causal event is given different evidence strength descriptions in the narrative text.

10. The method of claim 1, wherein, The spatio-temporal knowledge graph is constructed based on multi-source operation data reflecting the operation state of unmanned vehicles, which is obtained from at least a plurality of data sources including a scheduling system, a vehicle terminal device, and a work order system.

11. An analysis device for historical operation data, comprising: an extraction module configured to determine at least one analysis anchor point in a spatio-temporal knowledge graph representing the operation state of a vehicle fleet according to a traceability problem to be analyzed; a construction module configured to trace the associated relationship in the spatio-temporal knowledge graph from the at least one analysis anchor point to identify and construct a causal subgraph containing potential causal chains; a conversion module configured to convert the structured information of the causal subgraph into an analysis prompt for a large model; a reasoning module configured to obtain an analysis result of the traceability problem according to the reasoning of the large model on the analysis prompt.

12. An electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-10.

13. A non-transitory computer readable storage medium having stored thereon computer instructions, wherein, The computer instructions are used to make the computer execute the method according to any one of claims 1-10.

14. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any one of claims 1-10.

Images