Machine learning for predictive modeling in management of operations of EDM equipment product

To sustain and excel in competitive global market, organizations often bank on high productivity and world class quality. Endeavor of this research is to comprehend and model the manufacturing process of Electrical Discharge Machining (EDM) equipment product in order to increase productivity. Outcome of EDM operation is strongly influenced by various process parameters. The paper presents a framework based on machine learning algorithms to analyze the relationship between input process parameters and EDM response to build a predictive model of EDM operations. Physical experimentations have conducted considering Discharge Current, Pulse Duration, Duty Cycle and Discharge Voltage as independent variables while Material Removal Rate has been used as target variable. Four different machine learning algorithms namely Random Forest, Support Vector Regression, Elastic Net and Bagging have been adopted as applied predictive modeling tools. Results justify the usage of machine learning methods to deal with the research problem. Statistical analysis has been conducted as well for comparative performance analysis. Further correlation based supervised feature selection methodology has been applied to identify the key predictors.