Liquefaction prediction using rough set theory

Evaluation of liquefaction is one of the most important issues of geotechnical engineering. Liquefaction prediction depends on many factors, and the relationship between these factors is non-linear and complex. Different authors have proposed different methods for liquefaction prediction. These methods are mostly based on statistical approaches and neural network. In this paper, a new approach based on rough set data mining procedure is presented for liquefaction prediction. The rough set theory is a mathematical approach to the analysis of imperfect knowledge or unclear description of objects. In this approach, decision rules are derived from conditional attributes in rough set analysis, and the results are compared with actual field observations. The results of this study demonstrate that using this method can be helpful for liquefaction prediction and can reduce unnecessary costs in the site investigation process.