Publication

Abstract : Fuel Consumption (FC) is a major concern for the maritime industry, as ships are heavy consumers of fuel. Fuel cost is expensive, so predicting FC reduce the operating cost. A lot of traditional techniques predicted the ship's FC, but they still leave some problems such as limited accuracy under sea conditions, CO2 emission, and speed optimization. In order to overcome these issues, this study focuses on prediction of FC by ships using ML model. Here, the collected dataset that take input as CO2 emission, distance and engine efficiency to predict the FC with all the ML-based regression and classification models. These ML model are helpful in optimizing fuel usage through accurate prediction and monitoring of FC. Also, these chosen ML models are evaluated using performance evaluation metrics. Among all the ML models, the combined technique named Polynomial+CatBoost shows the best performance by achieving the minimum MAE value of 4.61.