Prediction of Machine Failure by Using Machine Learning Algorithm

Machine failure halt many processes and causes minimum usage of unexploited
resources. Prediction of the anomalies of a machine can act as an indicator and
precaution to avoid machine malfunction. Prior to that, the big data undergo
preprocessing; data transpose and imputation. Then, the data is cluster by using K
Means to produce labeled input that will be trained by using Gradient Boosting
Machine, a decision tree algorithm to make prediction. The columns consist of the
variables that record the reading of machine sensor tags. Validation for the model is
analyzed by using validation testing data and cross validation. Model built resulted in
variables importance’s ranking and subsequently, prediction can be made. The results
of the data analysis will be illustrated in a dashboard via Power BI. Consequently, the
user will be able to make an informed decision.