Enhanced Feature Selection Algorithm for Effective Supervised and Unsupervised Learning

Abstract

In machine learning applications, the dimension of input data is high with newlineredundancy and noisiness due to which model accuracy is highly affected in large data newlinesets. The model accuracy can be improved through dimensionality reduction by newlineminimizing the dimensions of original dataset and by eliminating redundant or newlineirrelevant information. Dimensionality reduction is realized by feature selection or newlinefeature extraction. newlineIn both supervised and unsupervised learning, many optimal feature newlineselection methods are used to find the most important variables or parameters which newlineare necessary in predicting the outcomes. Also feature selection helps to train the newlinealgorithm faster as well as reduces the model complexity. Although the number of newlineattributes available is more, only limited features will helps to enhance the accuracy of newlinethe model in both the types of learning newline