TRANSFER LEARNING OF A MACHINE-LEARNING MODEL USING A HYPERPARAMETER RESPONSE MODEL

摘要

The invention relates to a method of building a machine-learning model (301-304) for operational monitoring of an ensemble (200) of field devices (201-204), the method comprising: performing a training (2001, 2012) of a first value of a hyperparameter (2053) of the machine-learning model (301-304) based on training data (603) of a first device of the ensemble (200); and, based on the training: obtaining a response model (2050) of the hyperparameter, the response model providing a mapping of a value of the hyperparameter to a performance indicator of the machine-learning model; and determining a second value of the hyperparameter for a second field device (201-204) of the ensemble (200) based on the response model (2050).