CN-121998473-A - Dynamic image construction and intelligent management method and system for railway crew

CN121998473ACN 121998473 ACN121998473 ACN 121998473ACN-121998473-A

Abstract

The invention provides a dynamic image construction and intelligent management method and system for railway crew members, which relate to the technical field of computers and comprise the steps of obtaining multisource original data of the railway crew members; the method comprises the steps of carrying out data fusion according to multi-source original data, constructing an initial multiplication knowledge graph, carrying out feature aggregation according to the initial multiplication knowledge graph to obtain a standard multiplication knowledge graph, carrying out time sequence feature extraction according to the standard multiplication knowledge graph to obtain dynamic behavior features, carrying out matching and association analysis on the dynamic behavior features and loads and response conditions under different working scenes to obtain an image tag set, carrying out decision according to the image tag set, and positioning a training missing source by comparing real-time emergency treatment data with a preset threshold value and associating the knowledge graph to obtain an optimal management suggestion. The invention realizes dynamic and accurate portrait construction and intelligent management decision of crews.

Inventors

LIU JIANJUN
LI JIE
LI HAIBIN
ZHAO LIANG
ZHAO RUFENG
Song Qingshui
Xie Lukuan
LI JUN
WANG CHAOHUI
DING YI
WANG QI
WANG LINGYAN
YU JINJIN
LIU JIE
WANG PENG
ZHANG XIAO

Assignees

中国铁路北京局集团有限公司北京客运段

Dates

Publication Date: 20260508
Application Date: 20251215

Claims (10)

1. A dynamic portrait construction and intelligent management method facing railway crew members is characterized by comprising the following steps: acquiring multi-source original data of railway crew members, wherein the multi-source original data comprises service process data, passenger feedback data and personal qualification data; Carrying out data fusion according to the multi-source original data, and constructing and obtaining an initial multiplication knowledge graph by extracting four entities of crews, multiplication tasks, management events and examination item points and defining a cross-system semantic relation; Feature aggregation is carried out according to the initial multiplication knowledge graph, and a standard multiplication knowledge graph is obtained by calculating the attention weight between a value multiplication task and an installation management event and weighting and aggregating neighbor features; Extracting time sequence characteristics according to the standard passenger service knowledge graph, and capturing time sequence dependency relationship between service response and passenger transport event by analyzing dynamic evolution rule of service behavior data in time dimension to obtain dynamic behavior characteristics; Matching and association analysis are carried out on the dynamic behavior characteristics and the loads and response conditions under different working scenes, so that a portrait tag set is obtained; And making a decision according to the portrait tag set, and positioning a training missing source by comparing the real-time emergency treatment data with a preset threshold and correlating a knowledge graph to obtain an optimal management suggestion.
2. The method for constructing and intelligently managing dynamic images facing railway crews according to claim 1, wherein the method is characterized in that the method performs data fusion according to the multi-source original data, and the method constructs and obtains an initial crews knowledge graph by extracting four entities of crews, multiplication tasks, management events and examination item points and defining a cross-system semantic relation, and comprises the following steps: performing entity identification processing according to the multi-source original data, and identifying and extracting four types of entities and attributes thereof of the crew members, the multiplication tasks, the management events and the examination item points through analyzing the multiplication paths, the emergency treatment records and the examination item point evaluation data to obtain an entity set; performing relationship definition processing according to the entity set, and defining a cross-system semantic relationship among the entities based on a preset multiplication management service rule to obtain an initial relationship network; And carrying out map construction processing according to the initial relation network, and constructing to obtain an initial multiplication knowledge map by fusing the entity set and the initial relation network and carrying out completion and conflict resolution on entity attributes according to business rules.
3. The method for constructing and intelligently managing dynamic images for railway crew members according to claim 1, wherein feature aggregation is performed according to the initial crew knowledge graph, and the standard crew knowledge graph is obtained by calculating attention weights between a duty and an installation event and weighting and aggregating neighbor features, comprising: Performing feature projection processing according to the initial multiplication knowledge graph, and mapping entity features with different dimensions to a unified feature space by performing linear transformation on initial feature vectors of entities in the graph to obtain projected entity feature representations; Performing attention calculation processing according to the projected entity characteristic representation, and calculating attention weight coefficients between adjacent entities by analyzing service association strength between a multiplication task and an emergency disposal event to obtain association strength distribution between the entities; And carrying out feature weighted aggregation processing according to the association strength distribution, carrying out weighted summation on projection features of neighbor entities based on a preset attention weight coefficient, generating enhanced feature representation of each entity, and updating the feature of the map node to obtain a standard multiplication knowledge map.
4. The method for constructing and intelligently managing dynamic portraits for railway crew according to claim 1, wherein the method for extracting time sequence features according to the standard passenger knowledge graph, capturing time sequence dependency relationship between service response and passenger events by analyzing dynamic evolution rules of service behavior data in time dimension, and obtaining the dynamic behavior features comprises the following steps: carrying out time sequence data construction processing according to the standard passenger knowledge graph, and obtaining structured multidimensional time sequence data by extracting service response records and passenger event sequences associated with the crew members and aligning according to time windows of value times shift; Performing dynamic mode analysis processing according to the multidimensional time sequence data, and obtaining a time sequence characteristic mode by analyzing delay and continuous change rules of service response indexes relative to event trigger points in different passenger traffic scenes, wherein the time sequence characteristic mode comprises short-term fluctuation and long-term trend characteristics of service behaviors; And carrying out dependency modeling processing according to the time sequence characteristic mode, and capturing time sequence dependency between the service response characteristic and the passenger transport event characteristic by calculating the correlation of the service response characteristic and the passenger transport event characteristic under different time lags to obtain dynamic behavior characteristics.
5. The method for constructing and intelligently managing dynamic portraits for railway crew members according to claim 1, wherein the matching and association analysis of the dynamic behavior features with the load and response conditions under different working scenes is performed to obtain a portrayal tag set, comprising: Performing scene condition matching processing according to the dynamic behavior characteristics, and obtaining a preliminary scene matching result by matching service response time length and event handling frequency characteristics with a preset passenger flow load threshold value and an emergency handling standard; Carrying out multidimensional association analysis processing according to the preliminary scene matching result, and identifying the matching degree between service performance and service requirements by analyzing the association relation between the multiplication task type and the assessment item point score to obtain an association analysis result; and performing label generation processing according to the association analysis result, and generating a dynamic label representing service capacity of the crew member according to a preset label rule by integrating the scene matching result and the association analysis result to obtain a portrait label set.
6. A dynamic portrait construction and intelligent management system facing railway crew members is characterized by comprising: The acquisition module is used for acquiring multi-source original data of railway crews, wherein the multi-source original data comprises service process data, passenger feedback data and personal qualification data; the fusion module is used for carrying out data fusion according to the multi-source original data, and constructing and obtaining an initial multiplication knowledge graph by extracting four entities of crews, multiplication tasks, management events and examination item points and defining a cross-system semantic relation; the aggregation module is used for carrying out feature aggregation according to the initial multiplication knowledge graph, and obtaining a standard multiplication knowledge graph by calculating the attention weight between a valued multiplication task and an installation event and weighting and aggregating neighbor features; the extraction module is used for extracting time sequence characteristics according to the standard passenger knowledge graph, and capturing time sequence dependency relationship between service response and passenger events by analyzing dynamic evolution rules of service behavior data in time dimension to obtain dynamic behavior characteristics; the analysis module is used for carrying out matching and association analysis on the dynamic behavior characteristics and the loads and response conditions under different working scenes to obtain a portrait tag set; And the decision module is used for making a decision according to the portrait tag set, and obtaining an optimal management suggestion by comparing the real-time emergency treatment data with a preset threshold and associating a knowledge graph to locate a training missing source.
7. The railway crew oriented dynamic portrayal construction and intelligent management system of claim 6, wherein the fusion module comprises: The first fusion unit is used for carrying out entity identification processing according to the multi-source original data, and identifying and extracting four types of entities and attributes thereof of the crew members, the multiplication tasks, the management events and the assessment item points through analyzing the multiplication paths, the emergency treatment records and the assessment item point evaluation data to obtain an entity set; the second fusion unit is used for carrying out relation definition processing according to the entity set, defining a cross-system semantic relation among the entities based on a preset multiplication management service rule, and obtaining an initial relation network; and the third fusion unit is used for carrying out map construction processing according to the initial relation network, and carrying out complementation and conflict resolution on entity attributes according to business rules by fusing the entity set and the initial relation network to construct and obtain an initial multiplication knowledge map.
8. The railway crew oriented dynamic portrayal construction and intelligent management system of claim 6, wherein the aggregation module comprises: The first aggregation unit is used for carrying out feature projection processing according to the initial multiplication knowledge graph, and mapping entity features with different dimensions to a unified feature space through linear transformation of initial feature vectors of entities in the graph to obtain projected entity feature representations; The second aggregation unit is used for carrying out attention calculation processing according to the projected entity characteristic representation, and calculating attention weight coefficients between adjacent entities by analyzing service association strength between a multiplication task and an emergency disposal event to obtain association strength distribution between the entities; And the third polymerization unit is used for carrying out characteristic weighted aggregation processing according to the association strength distribution, carrying out weighted summation on the projection characteristics of the neighbor entities based on a preset attention weight coefficient, generating enhanced characteristic representation of each entity, and updating the node characteristics of the map to obtain a standard multiplication knowledge map.
9. The railway crew oriented dynamic portrayal construction and intelligent management system of claim 6, wherein the extraction module comprises: The first extraction unit is used for carrying out time sequence data construction processing according to the standard passenger knowledge graph, and obtaining structured multidimensional time sequence data by extracting service response records and passenger event sequences associated with the crew members and aligning according to time windows of value times shift; The second extraction unit is used for carrying out dynamic mode analysis processing according to the multidimensional time sequence data, and obtaining a time sequence characteristic mode by analyzing the delay and continuous change rules of service response indexes relative to event trigger points in different passenger traffic scenes, wherein the time sequence characteristic mode comprises short-term fluctuation and long-term trend characteristics of service behaviors; And the third extraction unit is used for carrying out dependency modeling processing according to the time sequence feature mode, and capturing the time sequence dependency between the service response feature and the passenger transport event feature by calculating the correlation of the service response feature and the passenger transport event feature under different time lags so as to obtain the dynamic behavior feature.
10. The railway crew oriented dynamic portrayal construction and intelligent management system of claim 6, wherein the analysis module comprises: The first analysis unit is used for carrying out scene condition matching processing according to the dynamic behavior characteristics, and obtaining a preliminary scene matching result by matching service response time length and event handling frequency characteristics with a preset passenger flow load threshold value and an emergency handling standard; The second analysis unit is used for carrying out multidimensional association analysis processing according to the preliminary scene matching result, and identifying the matching degree between the service performance and the service requirement by analyzing the association relation between the value multiplication task type and the assessment item point analysis to obtain an association analysis result; And the third analysis unit is used for carrying out label generation processing according to the association analysis result, and generating a dynamic label representing the service capacity of the crew member according to a preset label rule by integrating the scene matching result and the association analysis result to obtain a portrait label set.

Description

Dynamic image construction and intelligent management method and system for railway crew Technical Field The invention relates to the technical field of computers, in particular to a dynamic portrait construction and intelligent management method and system for railway crew members. Background In the field of intelligent management of railway transportation, the accurate depiction and evaluation of the working state and business capability of crew members are key links for improving the service quality and operation safety of passenger transportation, a traditional crowd portrayal technology is used for generating a labeled feature model by integrating multidimensional data, a basic framework is provided for personnel evaluation, in the railway passenger management practice, the prior art scheme mainly collects structured scheduling, attendance and basic assessment data through a single business system, and relies on preset static rules to carry out statistical analysis and simple classification on the data, so that limited static evaluation labels such as 'working efficiency', 'attendance compliance' and the like are generated, management decision is assisted in a form, however, the method is limited by the single and unstructured information of a data source and the lack of cross-system business association, deep perception of a complex dynamic operation scene is difficult to realize, an evolution rule and complex association between specific passenger transportation events and working loads on the basis of fixed rules and simple statistics cannot be captured, a portrayal result is static, a piece face and lag, and meanwhile, the prior art can provide accurate and precise response to the scheduling data in real-time dimension, and the problem of accurate positioning and the real-time performance can not be accurately interpreted, and the problem of the accurate positioning and the special service can not be provided in real-time, and the real-time performance can not be accurately interpreted, and the problem is difficult to provide. Based on the shortcomings of the prior art, a method and a system for constructing and intelligently managing dynamic images for railway crew members are needed. Disclosure of Invention The invention aims to provide a dynamic image construction and intelligent management method and system for railway crew members, so as to solve the problems. In order to achieve the above purpose, the technical scheme adopted by the invention is as follows: In a first aspect, the present application provides a dynamic portrait construction and intelligent management method for railway crew members, including: Acquiring multisource original data of railway crew members, wherein the multisource original data serve process data, passenger feedback data and personal qualification data; Carrying out data fusion according to the multi-source original data, and constructing and obtaining an initial multiplication knowledge graph by extracting four entities of crews, multiplication tasks, management events and examination item points and defining a cross-system semantic relation; Feature aggregation is carried out according to the initial multiplication knowledge graph, and a standard multiplication knowledge graph is obtained by calculating the attention weight between a value multiplication task and an installation management event and weighting and aggregating neighbor features; Extracting time sequence characteristics according to the standard passenger service knowledge graph, and capturing time sequence dependency relationship between service response and passenger transport event by analyzing dynamic evolution rule of service behavior data in time dimension to obtain dynamic behavior characteristics; Matching and association analysis are carried out on the dynamic behavior characteristics and the loads and response conditions under different working scenes, so that a portrait tag set is obtained; And making a decision according to the portrait tag set, and positioning a training missing source by comparing the real-time emergency treatment data with a preset threshold and correlating a knowledge graph to obtain an optimal management suggestion. In a second aspect, the present application also provides a dynamic portrait construction and intelligent management system for railway crew, including: The acquisition module is used for acquiring multi-source original data of railway crews, wherein the multi-source original data comprises service process data, passenger feedback data and personal qualification data; the fusion module is used for carrying out data fusion according to the multi-source original data, and constructing and obtaining an initial multiplication knowledge graph by extracting four entities of crews, multiplication tasks, management events and examination item points and defining a cross-system semantic relation; the aggregation module is used for carrying out feature aggregatio