A Machine Learning Based Sameness Recognition Method for Power System Management Information

State Grid Corporation's power construction project information comes from multiple regions and different scale power construction sub-project, and there are missing fillings and irregularities, which makes supervision difficult, and brings difficulties in project acceptance and management. Based on the random forest algorithm, our work proposes a sameness analysis method of engineering problem description text, and realizes the task of automatic classification of engineering defects. This work first defines the concept of sameness recognition in power construction project, and performs Chinese word segmentation on the problem description text in the engineering information data set after the standardized processing. Then, this work constructs a feature vector space. By implementing random forest algorithm, this work conducts sameness analysis and defect classification on the construction problem text, and proves the effect of the method proposed in this work through case analysis.