An Intelligent Expert System Based on Fuzzy Least squares Support Vector Machine for Gas Pipeline Safety Assessment

To solve the safety assessment of city underground gas pipeline, a novel approach to build an expert system was proposed, based on Fuzzy Least squares Support Vector Machine and the mathematical model. First, Support vector machines (SVMs) are introduced, which are learning algorithms derived from statistical learning theory. Then, the mathematical model is described, 8 factors affected the safety are selected through cluster analysis and correlation analysis. After that, we design the expert system architecture. Finally, the system is used practically in a city in China. The experimental result shows that our approach is validated with good generalization and robustness, which is better than BP nerve network.