Sharing a ride in vehicle is known as carpooling and many businesses are adapting this system. Carpooling is an effective way to control air pollution since it decreases the number of automobiles on busy roads. It also helps to reduce traffic roadways, allowing automobiles to drive faster. Carpooling reduces the pollution even further because vehicles use less gas when they travel without much stopping in the traffic. People may also drive their automobiles whenever they choose, with no need for cheaper maintenance and parking charges for each vehicle. According to studies, each car in a carpooling system takes up to 11 vehicles off the road. This cuts yearly pollution by 13 metric tonnes. In India, if carpooling done correctly, one single car will take around 25 automobiles off the road. To make carpooling easier, in this work, a machine learning (ML) model is utilized to identify the personality types of customers based on the tweets and to group them for the same car with similar personalities. The idea intends to break down this barrier between people who don't like to travel with the people who have different personality. In this model, Natural Language Processing (NLP) is used to deal with the dataset before passing it to the model to train. The proposed work is done using XG-Boost, Decision Tree Classifier, Support Vector Machine, and Stochastic Gradient Decent optimizer. Pre-processing the dataset employed the NLP techniques such as Tokenization, Lemmatization and Stemming. After evaluating all the algorithms, we found out that XG-Boost is outperforming all the remaining three algorithms by achieving 68% accuracy easily with 5-fold cross validation.
Machine Learning Based Personality Classification for Carpooling Application
2023-02-09
452174 byte
Conference paper
Electronic Resource
English
Advancements in Carpooling System using Machine Learning
IEEE | 2024
|Carpooling method and carpooling system based on trajectory analysis
European Patent Office | 2015
|Developing a Web-Based Carpooling Application
IEEE | 2023
|NTIS | 1990