The invention relates to the technical field of industrial economy analysis and power
data mining, and provides a power utilization
mode change-based industrial transfer implementation effect evaluation method, which comprises the following steps of: acquiring power utilization
time sequence data before and after transfer; extracting multi-dimensional features; the
relative change rate is calculated, and direction unification is carried out according to index'lifting / descending '; constructing a comprehensive effect index; performing grade judgment by adopting ordered judgment /
ensemble learning; and outputting a result and confidence. An
index system covering the dimensions of load, efficiency, stability and the like is constructed by establishing a power utilization mode
base line and comparison before and after transfer, and an effect quantification mechanism and an integrated learning model with unified directions are introduced, so that the industrial transfer effect is objectively and interpretably evaluated and graded; the confidence coefficient adopts geometric weighting of calibrated probability, stability, integrity and time span, and the method can be selectively combined with synthetic contrast / causal recognition to eliminate
hybrid influences of seasons, prices,
meteorology and the like, so that the evaluation robustness is improved.