Research on Evaluation Method for Potential Value of Comprehensive Energy Metering Data

In order to promote the value mining of comprehensive energy measurement data and improve the level of adaptation between data sets and business scenarios, research on evaluation methods for the potential value of comprehensive energy measurement data is carried out. Firstly, for different typical business scenarios, the Apriori algorithm is used to identify frequent itemsets of the multi-source data involved, and extract strong correlation characteristics to build a typical business scenario characteristic library. Then, the Fisher discriminant analysis method is used to calculate the contribution of each characteristic to the business goal, in order to reduce the modeling dimension and information redundancy, the Pearson correlation coefficient is used to obtain the correlation degree between each characteristic, and the high correlation characteristic is combined to form a super characteristic. Finally, a cost matrix is used to screen the characteristic combination with the highest degree of fitness and contribution based on the evaluation feedback of business objectives, which serves as the data characteristic set for modeling each business scenario. Experiments have shown that this method can capture unique information from various business scenarios and create a highly adaptable dataset for business objectives, transforming non quantifiable qualitative evaluation of "purpose" into quantifiable quantitative evaluation of "value", forming a scenario oriented data resource value estimation model, and providing technical support for promoting innovation in business formats based on multi-source data.