Feature selection using Linear Discriminant Analysis for breast cancer dataset

Breast cancer disease is one of the most common disease in women. Diagnosing this disease in earlier stage is very important. Generally for diagnosing this disease it takes too much of time because of more number of features (or) attributes that are passed to the classifiers. The classifiers have to run and analyze these features repeatedly as iteration and finally it gives the result. If the dataset has more number of attributes then it may take too much of time for diagnosing. Therefore the aim of this analysis is to reduce the variables and diagnose the disease effectively. In this experiment Wisconsin original dataset has been used and the features were passed to Linear Discriminant Analysis to analyze the performance of the attributes by reducing it. In this approach the features were reduced to four variables and these selected variables gave good accuracy of about 95% than other set of variables