Prediction of world happiness scenario effective in the period of COVID-19 pandemic, by artificial neuron network (ANN), support vector machine (SVM), and regression tree (RT)

Main Article Content

Sharma Khemraj , Phramaha Chakrapol Acharashubho Thepa , Hsinkuang Chi , Wann Yih Wu , Sasmita Samanta , Jyoti Prakash

Abstract

Objectives:To comparison, the effectiveness of prediction world happiness based on data of world happiness in theperiod of the COVIT-19 pandemic.
Methods:The prediction is used the qualitative analysis with further process of Artificial Neuron Network (ANN), Support Vector Machine (SVM), and Regression tree (RT) by MATLAB.
Results:The prediction of world happiness effective increasing of ladder weight that’s the trend of happiness based on data of six valuable impact of happiness consist of GDP per capita, Social support, Healthy lifeexpectancy, freedom to make life choices, Generosity, Corruption Perception, Residual error. The study found the significant ladder weight prediction increasingly used ANN, SVM, and regression tree analysis. The ANN # 3 model 6-20-1 is effective by the sign of the highest accuracy as 83.68%, among of diver error analysis i.e., SVM2 is found the significant testing RMSE0.5454, and the regression of Three2 model have significant testing of error RMSE 0.57815. The study found the prediction of World Happiness Effective Scenario on the period of COVIT-19 was found increasingly based on data of WHO from 2020consist with Finland from 7.809 (7.748-7.870) score in 2017-19 with 7.889 (7.784-7.995) score in 2020.
Conclusions:This is an advantage of SVM compared to neural networks, which have multiple solutions associated with local minima. Also, SVM can be used with short-length data set. Therefore, SVM is simpler to use than ANN but is the first choice of option for the world happiness prediction is ANN. Because the error analysis was suitable with the prediction the effectiveness of the world happiness rate of ladder and error values less.

Article Details

Section
Articles