Pengurangan Dimensi dengan Metode Linear Discriminant Analist (LDA)

Abstract

The purpose of this study is to reduce the dimensions of the dataset that affect the prediction of breast cancer. The data used in research is very much data or is called high-dimensional data. The use of classification algorithms has weaknesses when used on high-dimensional data, so an appropriate method is needed to reduce the dimensions or variables used. There are several methods that can be used to reduce dimensions. In this study using the method of linear discriminant analysis (LDA). LDA is a supervised machine learning algorithm that is used to classify data into several classes, using a linear technique to determine the best set of linear variables to unify class data. LDA is used to reduce the dataset variables used by retaining information that is important for the classification process. The method used in this research is using LDA in data processing and then using a logistic regression model for the classification process. The conclusion obtained in this study is that LDA can overcome the problem of multiclass classification. The results obtained were 16 wrong cases out of a total of 455 cases so that the results obtained were 0.035% misclassification.