Method and equipment for enhancing operation and maintenance abnormal data based on key performance indexes

The invention provides a key performance index-based operation and maintenance abnormal data enhancement method and equipment. The method comprises the following steps of: acquiring historical data; forming a matrix by the historical data; performing zero equalization on each row of the matrix; calculating a covariance matrix of the matrix after zero equalization; calculating a characteristic value of the covariance matrix; sorting the feature values, selecting a specific number of sorted feature values and feature vectors corresponding to the specific number of feature values for projection to obtain enhanced data, and inputting the enhanced data into a constructed network model for training; and classifying the key performance index data by using the trained network model to obtain abnormal data. According to the method, the dimension of the data sample is reduced, under the condition that the original data characteristics are guaranteed, the feature types are reduced, the collinearity among the data is eli