To improve the predicting precision and controlling precision of a learning device, and to successively execute a learning function by adding an exceptional input data discriminating function to the learning device by using a data similarity discriminating device.
The exceptionality of input data is discriminated by a data similarity discriminating device 103, and a processing corresponding to an exception is activated. Also, a model selecting device 105 selects a proper learning model from a model storing part 108 based on the discriminated result of the similarity data discriminating device 103. A mode updating device 104 operates the update processing of a learned model by using data stored in a model execution history storing part 113 and an inter-model subordinate relation storing part 114.
MAEDA AKIRA
MORITA TOYOHISA
OMORI KATSUMI
OISHI SATOSHI
ASHIDA HITOSHI
Next Patent: METHOD FOR PREDICTING ARRAYED SECONDARY STRUCTURE AND DEVICE THEREFOR