The Application Of Feature Selection On Triple Negative Breast Cancer

Triple Negative Breast Cancer is a kind of special breast cancer with ER, PR and HER2showing negative. Because of the incidence of TNBC is younger than any other breast cancers,the possibility of viscera’s transfer is much higher than any other breast cancers. And becauseof its poor prediction, there’s still no effective treatment on it. Chemotherapy is the only wayto cure TNBC. The neoadjuvant chemotherapy base on anthracyline is a best and populartherapeutic schedule on TNBC. The neoadjuvant chemotherapy have big breakout on TNBC.However, because of the sensitivity and resistivity of the drugs, a little part of TNBCbecomes sensitive to anthracycline, which affect the treatment. So it is for the best to test thedrugs’ sensitivity. But a little part of the TNBC has a strong sensitivity on anthracycline,which affect the treatment effect of the TNBC. If we could predictive what kind of person issensitive to anthracyline, it can improve the neoadjuvant chemotherapy’s popular. So we wantto find a model that could find out what kind of person is sensitivity to anthracyline. Thenimprove the treatment effect of TNBC.So the main idea of this paper as follows:1) Firstly, have knowledge of the current research status of TNBC and the basic therapeuticschedule, figure out that the key to improve the success rate of TNBC can be how topredict the TNBC patient whether sensitive to anthracycline.2) Introduce the basic procedure of feature selection and the classifier used in the text,prepare for the feature selection algorithm improvement and the feature selection ofTNBC.3) At the basic deep understanding of PCA and FLD algorithm, improve PCA and FLD,check if really improve the classification accuracy of datasets of SPECTF andionosphere.4) Use the improve PCA and FLD algorithm on select TNBC’s features to predict if theTNBC patient is sensitive to anthracycline. Summarizing and analysis the experimentconclusion.