Predictive Maintenance for Factory Equipment

Abstract

Predictive maintenance is a crucial strategy in
modern manufacturing that enables industries to
anticipate and prevent equipment failures, thereby
avoiding costly downtimes and improving operational
efficiency. This project proposes a machine learningbased
approach to predict factory equipment failure
using sensor data collected from machines. By
analyzing historical and real-time data through
supervised learning models such as SVM, Decision
Tree, Naïve Bayes, and Logistic Regression, the system
identifies early signs of malfunction. Among the
evaluated models, Support Vector Machine (SVM)
achieved the highest accuracy of 95%, demonstrating
its effectiveness for predictive maintenance tasks. The
solution leverages Python libraries and the Kaggle
Predictive Maintenance dataset for analysis,
visualization, and performance evaluation