Predictive Maintenance Using Machine Learning and Data Mining: A Pioneer Method Implemented to Greek Railways

In every business, the production of knowledge, coming from process effective information, is recognized as a strategic asset and source competitive advantage. field railways, vast amount data are produced, which necessary to be assessed, deployed in an optimum way, used mechanism, will lead making right decisions, aiming at saving resources maintain fundamental principle railways passengers’ safety. This paper uses stored-inactive Greek railway company, method mining applies machine learning techniques create decision support draw up risk control plan for trains. We make effort apply Machine Learning open software (Weka) obsolete procedures maintenance rolling stock company (hand-written work orders supervisors technicians, dealing with dysfunctions train unit by experience, lack planning coding malfunctions schedule). Using J48 M5P algorithms Weka software, recorded, processed, analyzed that can help monitor or discover, great accuracy, prevention possible damage stresses, without addition new recording devices—monitoring on trains, aim predicting diagnosis fleet. The innovative capable being tool optimization management’s performance trains provide appropriate information implementation technical ability order achieve greatest target importance