Machine tool machining abnormity monitoring method based on Bayesian probability learning under small sample

A machine tool machining abnormity monitoring method based on Bayesian probability learning under a small sample comprises the following steps: step A, collecting vibration information in a machine tool machining process; step B, feature extraction; step C, respectively establishing a training set and a test set for the feature samples; step D, probability estimation is carried out on the training set data samples; step E, constructing a multi-classification SVDD model; f, solving the multi-classification model constructed in the step 5; g, carrying out probability hypothesis on the hypersphere model; h, solving the model assumed in the step G; and step I, integrating an algorithm into a machine tool monitoring platform of a friend machine technology, and monitoring the whole process in the machining process in real time. According to the method, the anomaly detection precision and generalization ability are improved, random noise interference is effectively avoided, and the anomaly detection precision is fur