Accuracy Acquirement for Petrol Price Prediction Using Machine Learning Enhanced Random Forest Algorithm

Main Article Content

Vijay Anand M , Devi.T , N. Poornima, Gnanavel R


In day today life Worldwide the most trending analysis is on Petrol price deviations along forecasting analysis required for everyday activities. Production varies in each states and countries which needs analysis that helps to reduce the demand in fuels and gases. The vital and global need for transportation is fuel needs and on time availability in each and every place. Global positioning system (GPS) updated satellite locations for vehicle navigation to increase the required availability of fuels. Machine Learning (ML) an accurate technique to consume the fuel usage based on the travel distance and prediction using consumption of petrol. We propose a prediction model using random forest algorithm for statistical analysis and attribute based on the dataset collected. We have collected a petrol price variation data from bank bazar and gathered the price variation based on highest and lowest changes. From 2019 to 2021 march petrol price as highest and lowest at the end of year and its performance mentioned as Rise, Decrease, Stable, Unstable as an attributes for consumption range. Based on the vehicle’s size such as light weight for travel like cars, jeep and heavy weight vehicles such as container and Lorries the fluently changes takes place. According to the world availabilities of vehicles the traffic increased day by day, also decision making system also fluctuate based on the traffic.  Using random forest algorithm 87% of accuracy achieved and data attributes splits the data according to the decision making points and avoid overfitting to the values which identifies the Euclidean distance according to the independent parameters which produces accuracy in proposed model.

Article Details