Automobile fault code analysis method, storage medium and electronic equipment

The invention relates to an automobile fault code analysis method, a storage medium and electronic equipment, and the method comprises the steps: obtaining the data of a fault code common record, cleaning the data, and making a standardized label training set; model training is carried out, BERT is adopted as a training model, a cross entropy function or a focus loss function is adopted as a loss function, an Adam optimizer is adopted to calculate and update network parameters, and distributed training is carried out through a Pytorch framework; the trained model is used for analyzing an automobile fault code. According to the scheme, feature information points of automobile fault codes are learned by means of a powerful attention mechanism of the BERT model, decoding is carried out according to processed labels, automobile parts or subsystems which may have faults are recommended to a user, and an inspection range is designated. The automobile second-hand business can obtain the common meaning of the fault c