| Chinese domestic railway is all over the country, the density of which is great. As the barrier of the railway, the protection tree can absorb noise, isolate sand and reduce the adverse climate’s destructive effects on railways. But the maintenance of the protection tree and down-tree early warning are related to the safety of the railway. According to the needs of railway’s maintenance, downtree early warning system is practical technology, which is combined with hardware and software. Once the warning information obtained, the protection tree is treated technically to eliminate safety hazards.This article uses the decision tree algorithm of data mining technology. Using railway trees down a large number of historical data preprocessing department and railway dangerous tree trigger of training data set, find railway pound a rule, early warning information. In this paper, the main work is:(1) study the theoretical knowledge of data mining and decision tree algorithm and method of ID3 algorithm of decision tree building and its derivative of the characteristics of other algorithm description and comparison.(2) Characteristics of classical ID3 algorithm weighted improvement for FW(Feature Weight)- ID3. The algorithm for each attribute in railway pound a warning effect, the result of the pound for each attribute feature weighting. Given weight selection and optimization of the process of continuously, pound a warning system in railway more applicable to different geographical climate of railway maintenance work.(3) Railway pound a warning system design and implementation of the process. The railway departments to provide data preprocessing, determine the selection properties, use the FW- create a decision tree ID3 algorithm, the rules are applied to railway pound a warning system. Through the simulation test of classical ID3 algorithm and FW- ID3 algorithm of two kinds of decision tree to predict performance, compare the result of the forecast. |
PDF Full Text Request