Open AccessImportant Feature Selection for Predicting Human Freedom Index Score using Machine Learning Algorithms
Author(s) -
Preeti Kumari,
S Prasad Babu Vagolu,
Sunil Chandolu
Publication year - 2020
Publication title -
international journal of innovative technology and exploring engineering
Language(s) - English
Resource type - Journals
ISSN - 2278-3075
DOI - 10.35940/ijitee.f3881.049620
Subject(s) - overfitting , artificial intelligence , degrees of freedom (physics and chemistry) , computer science , machine learning , entropy (arrow of time) , index (typography) , measure (data warehouse) , economic freedom , feature selection , index of economic freedom , feature (linguistics) , data mining , artificial neural network , law , physics , quantum mechanics , world wide web , political science , linguistics , philosophy
Human freedom index refers to the state of human freedom in various countries based their personal and economic attributes. Human freedom can help us identify nobility of citizens in a country. For an individual of a country freedom is of great value and hence it is worthy to measure. Though there are many attributes to measure the human freedom index both in personal as well as in economic factors, here we are interested to find only those features which contribute the most and are relevant to predict the outcome i.e. human freedom index score. We will go through various features engineering process like removing strongly correlated attributes, filtering method using Mutual Information (Entropy) and then use Select KBest algorithm to select top features filtered through Mutual information. These steps will help reduce the training time, increase accuracy and reduce overfitting when model is created to predict the human freedom index score which is a Machine Learning Regression problem.