Comparing Machine Learning Models For Predicting Fuel Consumption In Energy Generation For The Food Processing Industry In Nigeria
Authors
Ofili, Chiamaka Praise
Issue Date
2023-08
Degree
Publisher
Dublin Business School
Rights holder
Rights
Items in eSource are protected by copyright. Previously published items are made available in accordance with the copyright policy of the publisher/copyright holder.
Abstract
Accurate prediction of fuel consumption is crucial for effective energy management, cost optimization and environmental pollution management in the food processing industry in Nigeria. Machine learning models have shown promise in various domains for predicting and optimizing fuel and energy consumption. This research under the CRISP-DM methodology, compared the performance of three machine learning models which include Artificial Neural Networks, Support Vector Machines and Random Forest in predicting fuel consumption in energy generation in a food processing industry in Nigeria.
This research was conducted using historical data on fuel consumption and independent variables such as energy generated, gas consumption, gas pressure, etc. collected over 2 years from Flour Mills of Nigeria Plc.
The performance of the predicted models was evaluated based on Root Mean Square Error (RMSE), Mean Square Error (MSE) and R Square evaluation metrics. The Random Forest Model performed best across most metrics, with the lowest RMSE (8124.62 & 26061.43) and MSE (6.6E+07 & 6.8E+08) on both training and testing data, and a relatively high R-squared value (0.93 & 0.52). The ANN Model performed reasonably well, but the SVM Model had a comparatively poorer performance.