Energy efficiency optimization working condition recommendation method based on multi-source data model

The invention provides an energy efficiency optimization working condition recommendation method based on a multi-source data model. The energy efficiency optimization working condition recommendation method comprises the steps of 1, obtaining sample data; 2, initializing parameters; step 3, data preprocessing; 4, merging with analysis data; 5, historical pre-training is carried out; step 6, an incremental data training mode; step 7, determining sensitive features; step 8, machine learning analysis; and 9, outputting a result. According to the method, an energy efficiency optimization framework process based on multi-modal incremental data fusion is established, and the situation that an analysis result deviates from the reality due to incomplete measurement point arrangement and single analysis data is avoided. According to the method, in view of imbalance of unit equipment operation condition data, energy efficiency optimization targets under different working conditions are predicted through sensitive vari