A study on the use of machine learning methods for incidence prediction in high-speed train tracks

摘要: In this paper a study of the application methods based on Computational Intelligence (CI) procedures to forecasting problem in railway maintenance is presented. Railway an important and long-standing that critical for safe, comfortable economic transportation. With advent high-speed lines, has even more importance nowadays. We have developed study, applying from Statistics CI, examine feasibility predicting one-month-ahead faults two lines Spain. The data are recorded by measurement train which traverses monthly. results indicate CI competitive task against Statistical regression methods, with e-support vector outperforming other employed methods. So, feasible it useful planning process track maintenance.