Machine Failure Prediction using Supervised Machine Learning Technique

We focus on machine failure prediction in industry 4.0.Indeed, it is used for classification problems on the reliability and quality of their machines and products. We compare machine learning methods applied to a difficult real-world problem: predicting machine failure using attributes monitored internally by individual parts. The problem is one of detecting rare events in a time series of noisy and non-parametrically-distributed data. We develop a new algorithm based on the multiple-instance learning framework and the Regression algorithm which is specifically designed for the classification problems, and is shown to have promising performance. Its implementation is modular and extensible to support changes in the underlying production processes and the gathered data. It involves; loading, exploratory data analysis, training and model evaluation. The primary algorithms used in the project is Logistic regression algorithm. It is predictive analysis that describes data and explains the relationship between variables. Our results suggest that nonparametric statistical tests should be considered for learning problems involving detecting rare events in time series data. As large-scale systems continue to grow in scale and complexity, mitigating the impact of failure and providing accurate predictions with sufficient lead time remains a challenging research problem. Developing an accurate failure prediction model requires a critical understanding of the characteristics of real system failures. Experimental results indicates that the average prediction accuracy of our model using Logistic regression algorithm when failure is 90% accurate. The best performance overall was achieved with Logistic regression algorithm, although computational times were much longer and there were many more parameters to set