Application Research Of Decision Tree Algorithm In Engineering Quality Supervision Decision Support System

“Quality is greater than the word of life. Responsibility is heavier than a symbol of great weight”. The quality has always been the focus of attention in the engineering construction. Quality supervision is an effective means to ensure the quality of the project. Traditional engineering quality supervision is directed mainly at business design. While the project quality supervision focuses on decision support is relatively ignored. Large amounts of data have been stored during years of project quality supervision. It has been an urgent need to be solved that how to provide decision support for supervision department using analysis and prediction based on amounts of historical data.A system targeting at engineering quality supervision and decision support is designed using data mining and JAVA development. It is developed based on both decision support system condition in China and abroad, and the situation of Guizhou engineering quality supervision.The research is based on great references and technical documents and uses the barrage project evaluation data as a foundation of analysis. With the null value and outlier elimination and attribute reduction operation, data preprocessing is applied in the first place. A model is created using C4.5 algorithm based on WEKA platform then. Targeting at cost increasing problem caused by C4.5 calculation method(pruning after building), a modified algorithm is provided. This developed algorithm merges the over-fitting branches early during decision tree construction. It solves the problem of decision tree overloading and modeling time consuming. Experiment results illuminate that modeling accuracy and reduction are perceptibly improved by this algorithm. Improved algorithm is finally merged into WEKA platform through the second development. And a prediction standard is concluded based on the decision tree model. This prediction standard is applied in ?Engineering quality supervision and decision support system?. It provides support for quality inspection department.In this research, importance degree of different subprojects is pointed out through engineering quality supervision data analysis. It allows inspection department to apply more targeting quality supervision Also, it will predict potential problem in future project thus reduce probability of engineering accident.