Analyzing Ola Data for Predicting Price Based Trip Distance Using Random Forest and Linear Regression Analysis

Abstract

The paper aims to create a most efficient and accurate cab fare prediction system using machine learning algorithms and comparing them. The machine learning algorithms are Random forest algorithm and Linear regression and comparing the r-square, mean square error (MSE), Root MSE and Root Mean Squared Logarithmic Error (RMSLE) values. We implement the Random forest and linear regression algorithms to predict the prices of the system and to get the best accuracy when comparing both the algorithms. The algorithms should be efficient to predict the prices of the trips before the starting of the trip. The sample size considered for this work is N=10 for each of the groups considered. Totally it was iterated 20 times for efficient and accurate analysis on prediction of price with G-power in 80% and threshold 0.05%, CI 95% mean and standard deviation. The sample size calculation was done with clincle. Based on the statistical analysis the significance value for calculating the r-square was found to be 0.034. The Random forest algorithm gives a slightly better accuracy rate with a mean r-square percentage of 71.67% and the linear regression has mean r-square value of 70.57%. By this process, the prediction is done for the price prediction of the online cab rental system and the Random forest algorithm gives a better r-square value compared to the Linear regression algorithm.