Reza Rasinojehdehi; Goran Cirovic
Abstract
In recent years, there has been a surge in research exploring the potential of Machine Learning (ML) for predicting water pump failures. While some studies have focused on supervised approaches, others have delved into unsupervised methods. However, the challenge lies in identifying the key variables ...
Read More
In recent years, there has been a surge in research exploring the potential of Machine Learning (ML) for predicting water pump failures. While some studies have focused on supervised approaches, others have delved into unsupervised methods. However, the challenge lies in identifying the key variables crucial for accurate failure predictions. This study bridges this gap by consulting domain experts to discern essential variables, including water catchment area level, water quality index, lubrication frequency, water reservoir temperature, operating time, and power interruptions count. Employing supervised ML methods, specifically multiple regression and decision tree cart, the research aims to enhance the precision of failure predictions, shedding light on less-explored variables that play a significant role in pump failure.
