Automobile air conditioner fault prediction method and system based on data analysis and medium

• Main Authors: ZHU SONGLIN, WEN CHANGSHENG
• Format: Patent
• Language: Chinese; English
• Published: 04.06.2024
• Subjects: ARRANGEMENTS OR ADAPTATIONS OF HEATING, COOLING, VENTILATING,OR OTHER AIR-TREATING DEVICES SPECIALLY FOR PASSENGER OR GOODSSPACES OF VEHICLES; CALCULATING; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PERFORMING OPERATIONS; PHYSICS; TRANSPORTING; VEHICLES IN GENERAL

The invention discloses an automobile air conditioner fault prediction method and system based on data analysis and a medium, and belongs to the field of data mining. The method comprises the following steps: acquiring cleaning data of a plurality of automobile air conditioners and corresponding related information, and calculating use time; fault feature vectors are constructed, features comprise air conditioner model information, environment information, mileage and use time, the air conditioner model information and the environment information are numbered, and the mileage and the use time are normalized; the numbered air conditioner model information, the numbered environment information and the normalized driving mileage serve as input features of a use time prediction model, the use time serves as a prediction value, and a final use time prediction model is obtained through training of a use time gradient decision tree classification model; and the predicted use time of the to-be-tested air conditioner