Machine learning methods and machine learning device for learning error conditions, and failure prediction apparatus and failure prediction system including the machine learning device

A failure prediction system 1 includes a machine learning means 5 learns the conditions in connection with an error of an industrial machine. 2 The machine learning device includes a state monitoring unit which, during the industrial machine 2 is in operation or at rest, observed a state variable, which includes, for example, a data output of a sensor, internal data of a control software or computational data obtained based on these data, a determining data-receiving unit receives the determination data used to determine whether an error in the industrial machine 2 has occurred or to determine the degree of error, and a learning unit, which learns the conditions in the failure of the industrial machine 2 are related, according to a training data set is generated based on a combination of state variables with the determination data.