Happiness index prediction using machine learning algorithms

Abstract

For many years, individual welfare and world happiness are an important research point in all over the world. The predictor of life quality is connected closely with the world happiness index. Economic output, social support, life expectancy, freedom, the absence of corruption, and generosity are six major variables that can affect the global happiness index. This paper presents two machine learning methods to analyze the dataset, which is supported vector machine (SVM) and nave bayes. Then, the authors used them to predict the next year's happiness index by observing the images that the authors concluded from these two methods. Thus, the authors can get a clearer picture of peoples life quality as well. In the first step, the authors first perform feature normalization and input the features into the two algorithms. After that, the authors found that SVM was able to achieve better results with 92% and Nave Bayes with 87%. In addition, the authors analyze the significance of the indicators and the authors find that the factors that most affect the happiness index of the country are economic and medical. Those factors are very important things for each country. Moreover, before starting the analysis, the authors made some predictions. In our point of views, the economic, health, and social support should be the largest effect on happiness.