Dimensionality Reduction based on UncertainGraph Model

Classification in machine learning is done bymany factors which called attributes. The higher the numberof features, the more difficult it becomes to visualize thetraining set and then work on it. Sometimes, most of thesefeatures are related to each other and are therefore consideredredundant features. This is where Dimensionality Reduction(DR) algorithms come into play. In machine learning andstatistics, dimensionality reduction is the process of reducingthe number of supervised random variables by obtaining a setof main variables. Dimensionality reduction can be divided intofeature selection and feature extraction. This paper proposes anew Dimensionality reduction algorithm in the featureselection category using Pearson correlation of attributes andmaking uncertain graph models. The proposed model can bedone for any number of features with increasing theclassification performance compared with filter and wrapperstrategies

Keywords:

Dimensionality Reduction (DR) , Classification , Attributes , Feature Selection.