Multi-model data analysis method and system based on non-intrusive load monitoring

The invention relates to the technical field of non-intrusive load monitoring, and provides a multi-model data analysis method and system based on non-intrusive load monitoring, and the method comprises the steps: screening a part of data from power utilization data as a training set I; specifying a device label and a transient feature in the training set I, and calculating a first feature function of the training set I; constructing a decision tree model according to the first feature function and the transient feature; detecting the decision tree model; outputting a final decision tree model according to a detection result; classifying the total power utilization power data of the sample users through the final decision tree model; inputting the classified total power data into a deep learning model for training to obtain a non-intrusive deep learning model; respectively inputting actual power utilization total power data into the two models so as to obtain a power dynamic sequence of single electric equipm