A recommender subsystem construction for calculating the probability of a violation by a locomotive driver using machine-learning algorithms
This article describes the issues of analysis and assessment of the human factor for predicting the violation committed by the locomotive driver when driving the electric rolling stock. An intelligent system overview for assessing the likelihood of a violation by a locomotive driver is given. Such a system can generate recommendations depending on previously committed violations. One of the tasks is to reduce the risk of locomotive safety devices malfunctions, which are part of the locomotive electrical equipment. The solution to the problem of predicting the occurrence of possible violations is solved using tools and machine learning algorithms. A model has been built that generates recommendations for the driver based on information about previously committed violations and several static characteristics of the locomotive driver.