A DATA MINING FRAMEWORK USING KMEANS TO ANALYZE HIGHWAY ACCIDENT DATA IN MAHARASHTRA

Data mining has been a consistent technique to analyze road traffic accidents (RTA) for productive results. Sometimes, it is found that RTA occurrences are more frequent at certain road features. The accident analysis of these road features can help in identifying certain road accident characteristics that make a road accident to occur frequently in these road conditions. The objective of this work is to find out correlations of road features with other attributes that lead to an accident of a section of National Highway4 NHDP Phase-V. This part of Highway traverses through a plain as well as hilly terrains. The study is for a six-lane road between chainage 725.000 (Anewadi) to 865.350 (Khed-shivapur) to identify some additional safety measures to curb RTA. Databases of NHAI, Pune regarding Road traffic Accidents (RTA) were explored to compile this article. A systematic investigation will provide dynamic understanding about the highway accidents, incidence and causes. The predictive ability is perceived using cluster analysis along with the Apriori algorithm and the Confusion matrix. The validation tools were applied to test the ability of models to predict accidents.