Abnormal temperature detection device, abnormal temperature detection method, and non-transitory computer readable storage medium

An abnormal temperature detection method comprising:

• acquiring temperature data by a temperature acquirer;
• performing, by a preprocessor, preprocessing of calculating a feature amount of the temperature data;
• performing, by a synthesizer, synthesis processing of synthesizing the feature amount and past determination information indicating whether the temperature data acquired in the past is normal or abnormal;
• generating, by a learner, an identification model by performing machine learning on a result of synthesis by the synthesizer;
• acquiring temperature data for detection by the temperature acquirer;
• performing, by the preprocessor, the preprocessing on the temperature data for detection;
• performing, by the synthesizer, the synthesis processing by using a result of the preprocessing by the preprocessor;
• and
• determining, by a determiner, whether the temperature data for detection on which the preprocessing and the synthesis processing are performed is normal or abnormal by using the identification model.

A non-transitory computer readable storage medium storing a program executed by a computer, the program instructing the computer to:

• acquire temperature data;
• perform preprocessing of calculating a feature amount of the temperature data;
• perform synthesis processing of synthesizing the feature amount and past determination information indicating whether the temperature data acquired in the past is normal or abnormal;
• generate an identification model by performing machine learning on a result of the synthesis processing;
• acquire temperature data for detection;
• perform the preprocessing on the temperature data for detection;
• perform the synthesis processing by using a result of the preprocessing;
• and
• determine whether the temperature data for detection on which the preprocessing and the synthesis processing are performed is normal or abnormal by using the identification model.