Abstract This paper proposes an intelligent assessment method for the quality of power load datasets, based on large models and prompt engineering for quality assessment and automatic generation of quality assessment reports. By evaluating the completeness, accuracy, consistency, and temporality of the datasets, it identifies existing missing values, interference items, and temporality issues. Experiments show that this method can efficiently and automatically identify outliers and missing data, improving the accuracy and consistency of data quality assessment. The accuracy of outlier detection reaches 85%, and it enables the automatic generation of quality assessment reports. The research results have broad application potential in the quality assessment of power datasets and related fields.
Xie et al. (Fri,) studied this question.