Improvement of MTS Based on Rough Set Theory and Its Application in Classification
Lin Zh · Shuxue de shijian yu renshi · 2015
MTS is a method for pattern recognition which was first proposed by Genichi Taguchi.MTS integrates orthogonal test and signal-noise ratio with MD to select variables,so we can diagnose and evaluate target groups or make predictions.Considering that MTS may have some shortcomings in selecting variables and RS is good at dealing with it even when the information is uncertain,we use RS to select variables to improve MTS.Early detection of cancer cells' is helpful to the prevention and treatment of breast cancer.With the background of cancer cells' classification and detection,We select 600 cells from the UCI database as the research sample and use the improved MTS method to distinguish the cancer cells from the normal ones.The study shows that improved MTS has higher classification accuracy than MTS.The decrease of variables' s number due to RS can greatly simplify data collection work.This improved MTS method can provide technical reference to the detection of breast cancer and will be greatly helpful to other classification issues.