Research on the quality risk assessment model for water conservancy projects based on the semi-supervised classification of text

Abstract

Abstract This paper systematically analyzes the quality supervision data system of water conservancy projects. Then, according to the characteristics of water conservancy project quality supervision text, the Word2vec algorithm and TFIDF algorithm are combined to construct a feature extraction system of water conservancy project quality supervision text suitable for short length and few samples. Finally, a semi-supervisory model system consisting of logical regression, simple Bayes, and SVM is introduced to solve the problem of incomplete quality supervision risk data for water conservancy projects. To sum up, on the basis of the three parts, i.e. data system, feature extraction, and semi-supervisory text classification, we build a water conservancy project quality risk assessment framework and provide a data processing tool for machine learning of hydraulic engineering quality risks.