Efficient Analysis of Road Accident using Random Forest Comparison over Decision Tree Algorithm to Improve Accuracy to Predict the Severity of Accident

Abstract

Aim: To analyze road accidents using Random Forest comparison over Decision Tree algorithms to improve the accuracy and  to predict the severity of the accidents. Materials and Methods: Random Forest Algorithm and Decision Tree algorithm used to predict the accuracy percentage of Traffic accident severity analysis. The sample size was measured as 10.  G-power is calculated for two different groups, alpha (0.05), power (80%). Here the sample size is 40 and the iteration of the algorithms is 30, to attain the significant value is 0.02. Results: The results achieved with p=0.02 shows that two groups are statistically significant. It was observed that the Random Forest algorithm obtains the accuracy as 85.80. It appears to have better accuracy than the Decision Tree  algorithm 78.80%. Conclusion: This study concludes that the Random Forest Algorithm 85.80% performs significantly better than Decision Tree 78.80%.