Font Size: a A A

Research On Railway Accident Data Mining And Forecasting And Early Warning Method Based On The Correlation Analysis

Posted on:2017-04-21Degree:MasterType:Thesis
Country:ChinaCandidate:K N WangFull Text:PDF
GTID:2272330482987143Subject:Control Science and Engineering
Abstract/Summary:PDF Full Text Request
China’s railway transportation network has a long transport mileage, through multi-site, a wide range of radiation, large transport turnover, fast transport speed, etc. Annual passenger turnover and cargo turnover ranks first in the world.Railway safety covers a wide range, highly specialized and complex, along with the improvement of the railtransport speed and traffic density the risk has increased, the possibility of different nature of accidents also increased. Once the railway accident happens, not only cause casualties, economic losses and environmental damage, but also cause serious social impact. Therefore, research scientific for railway accident data and give full play to its role in decision support work has a very important significance for railway safety and stable development of China’s railway system.In this paper, the data start from the actual railway accidents. First, summarize all factors which affect the railway accident statistics, remove redundant ambiguities, and get in-depth analysis of the data samples available. Second, the association rules analysis of actual data is applied in railway accident studies combined with emerging datamining techniques and use the classical algorithm Apriori method study sampledata. After screening the most prominent of the strong association rules, the rules of data fields divided into specific safety hazard content and nature of the accident category, which guided by the Gray Relevance Theory. Through quantitative analysis of gray correlation degree calculating the intrinsic link between the two, further analyze the calculation results. Then, use the combination of forecasting principles based on the GM(1,1) and prediction value of cubic exponential smoothing, combined with the grey theory and the concept of IOWA operator, constructing a scientific and rational combination model. Predict respectively different nature of rail accidents and check errors. Last, summarize data situation and analyze security level of all units report, combined hazards and accidents to determine the severity of the relationship. Warn the accident which exceeds the value of the security, so analyze and warn early for railway accidents on China’s security.Full text based on actual accidents data, conduct quantitative research on various factors and interrelation, form safety warning application based on security risks and accident nature, which has a great significance to improve the railway safety level.
Keywords/Search Tags:Railway accident, Data mining, Correlation analysis, Combination forecasting, Security early warning
PDF Full Text Request
Related items