Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Railway accident cause analysis method based on word extension LDA

A technology of accident causes and analysis methods, applied in the fields of resources, instruments, electrical and digital data processing, etc., can solve problems such as the decline of expert judgment ability and the influence of subjective accident analysis results, and achieve the effect of deepening understanding

Active Publication Date: 2019-11-19
BEIJING JIAOTONG UNIV
View PDF7 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

A large amount of repetitive work will reduce the judgment ability of experts, and the subjectivity of expert judgment will affect the results of accident analysis

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Railway accident cause analysis method based on word extension LDA
  • Railway accident cause analysis method based on word extension LDA
  • Railway accident cause analysis method based on word extension LDA

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0046] like figure 1 As shown, the first embodiment of the present invention provides a method for analyzing the causes of railway accidents, and the method includes the following steps:

[0047] Step S110: use TF-IDF to represent the text of the railway accident, build a document vector space model, and generate a document vector;

[0048] Step S120: using the TextRank method to calculate the importance of words in the railway accident text;

[0049] Step S130: according to the word importance and semantic similarity, weight the words that meet the semantic threshold, and train the generated word extended LDA model;

[0050] Step S140: using the word extension LDA model to extract the text features of the railway accident, and extracting the theme and feature items of the cause of the railway accident;

[0051] Step S150: Use the SVM accident classification model to classify the text of the railway accident report, and determine the railway accident cause data set;

[0052...

Embodiment 2

[0064] like figure 2 As shown, the second embodiment of the present invention provides a word extension LDA topic model construction method based on word importance and semantic similarity, and the method includes the following process steps:

[0065] Step 1.1. Use the TextRank method to calculate the importance of words in the document

[0066] Specifically, the given accident text is divided into complete sentences, word segmentation is performed for each sentence, and stop words are removed, each sentence is represented as a set of phrases, a word graph is constructed, and then a co-occurrence relationship is used to construct any The edge between two words, only when two words co-occur in a fixed-length window, there is an edge between them, the importance of all words is initialized, and the importance of each word is calculated through multiple iterations, by setting The maximum number of iterations is used to control the calculation, and the final iteration result is ...

Embodiment 3

[0076] like image 3 As shown, the third embodiment of the present invention provides a text classification method based on two-level accident causes using the SVM accident classification model, and the method includes the following steps:

[0077] Step 2.1. Build an improved HFACS-RAs model.

[0078] Using the word-expanded LDA topic model generated by the training in Example 2 to extract the topic features of the accident text, and each topic selects the top eight topic words in the frequency ranking as the accident causative feature items to form the accident causation feature space; from the meaning of the topic words The human factor and organizational classification in the current accident can be identified. Based on the content extracted from the accident text features, an improved HFACS-RAs model is designed on the basis of the HFACS-RAs model, such as Figure 4 As shown, "preconditions for unsafe behavior" are further divided into "personal conditions for unsafe beha...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention provides a railway accident cause analysis method based on word extension LDA, and relates to the technical field of railway train operation safety analysis, and the method comprises thesteps: extracting an accident cause related theme and a theme word through employing a word extension LDA model based on the text content of a railway accident report; classifying the theme feature items according to a human factor and system classification method HFACS to form an improved HFACS-RAs model; performing text classification on the accident report by utilizing an SVM, and determiningan accident cause data set; optimizing bayesian network structure by combining chi-square test with unconstrained 0 / 1 optimization; estimating a CPT parameter of the Bayesian network by using a Logistic regression model; and determining an accident cause analysis model based on the improved Bayesian network, and calculating to obtain a key cause causing an accident result. Accident cause feature extraction is completed based on the word expansion LDA model, factors causing railway accidents and the influence degree of the factors on accident consequences are determined, understanding of the accident process is deepened, and measures are taken to prevent similar accidents from happening again.

Description

technical field [0001] The invention relates to the technical field of railway train operation safety analysis, in particular to a method for analyzing the causes of railway accidents based on word expansion LDA. Background technique [0002] It is of great significance to analyze the human factors and organizational factors in railway accidents. After the accident, in order to take effective preventive measures to prevent the recurrence of similar accidents, it is very important to diagnose and locate the root cause of the accident. The analysis of accident causes in my country's railway industry started relatively late. At present, for the research on the cause analysis method of railway accidents, or using the algorithm combining dissipative structure and entropy theory to study the evolution mechanism of high-speed railway operation accidents; or using the behavioral safety "2-4" model to analyze the root, fundamental, indirect, and direct causes of accidents The cause ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06F17/27G06K9/62G06Q10/06G06Q50/30
CPCG06Q10/0635G06F18/2411G06Q50/40
Inventor 张国琛郑伟
Owner BEIJING JIAOTONG UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com