Published Paper Details:

Estimating taxi fares has been an important research field within the transport industry since it homeworks pricing trends and transparency of the system both for service providers and customers. The project aims to find machine learning models that predict taxi trip fares using numerous variables, such as distance travelled during trips, time of the day, traffic conditions, number of passengers, and weather. Fare estimate will be addressed with robust data preprocessing, optimal feature engineering, and advanced model training using a synthetic dataset "constructed for practical regression tasks. The dataset is a rich source of over a 1000 data points with value-key trips varying the trips duration and fare amount, as well as contextual parameters like traffic and weather. This dataset will provide you with real world problems like missing values, outliers, correlation of features all together in one bundle. Applying and comparing ML models like Random Forest and Logistic regression and decision tree on this dataset based on Realization above had proven that Random forest Model proved the best with lower values in MSE, and was capable of fitting even non-linear relationship between the features. Along with its machine-learning train and evaluation functionality, it also provides a lightweight mechanism for predicting fares from input at application runtime. It also aims to collect the data on traffic, updates from weather and real-life datasets which could to be placed into next architecture to enhance the data feeding and to improve the model adaptability in future attempts. The predictive analytics embodied in this work speaks to power as it pertains to the taxi domain in such a way where it burns stronger in the empirical sense given the scope of paradigm machine in the usage to enhance fare predictions and decision making in transportation's dynamic environment.