Authors : Geetha Ramani and Shanthi Selvaraj
Abstract: Road accident analysis is very challenging task and investigating the dependencies between the attributes become complex because of many environmental and road related factors. Use of feature selection algorithms such as feature ranking, Fisher’s test, CFS etc which are using chi square test to find the statistical significance level of features is prevalent in data mining applications but it has been noted that such arbitrary usage may not be appropriate always. While analyzing the available data, sparse information may also be an issue wherein the study tends to use an option to probe the associations more rationally using Random Effect Model (REM). This approach could be much useful in understanding the rules of statistical significance that may be shadowed in other methods, especially chi square analysis. Results have shown the possible association together with amount of heterogeneity between the important variables involved in a data set related to road accidents that have occurred in Coimbatore, Tamil Nadu, India. Further, comprehensive analytical algorithms that are implemented in R have been provided to facilitate the approach for future replications.
Geetha Ramani and Shanthi Selvaraj, 2016. A Novel Approach to Analyze a Combination of IxJ Categorical Data for Estimating Road Accident Risk. Asian Journal of Information Technology, 15: 2005-2015.