Data feature classifying method based on feature extraction algorithm

The invention belongs to the technical field of clinical data detection and data mining, and particularly relates to a data feature classifying method based on the feature extraction algorithm. The method comprises the steps of collecting and establishing the vital sign database N of a normal person; collecting vital sign data S-1 of a patient; selecting the vital sign data N-1 of the normal person matched with the vital sign data S-1 of the patient from the database N, and extracting the feature with the highest resolving ability from the time feature sequence of the vital sign data S-1 of the patient with the feature extraction algorithm; selecting corresponding feature subsets from all extracted features; classifying the selected feature subsets with the data mining algorithm. According to the method, by mining real-time and high-frequency patient vital sign data detected by wireless medical equipment, clinical deterioration of the patient is prevented, early warning of death is provided, applicability is high, and working efficiency is improved.