Machine Learning Based Multi-Variable Happiness Index Prediction Model
Loading...
Date
2024
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
UMT, Lahore
Abstract
Happiness index is one of the new way to understand the happiness quantitatively. The
introduction of the happiness report in 2011 and its availability afterward every year
have given the happiness more scope to be studied and to be further understood as well.
The key goal of the thesis was to develop and suggest a machine learning based model
which can be further utilized, in the prediction of Happiness Index. The datasets used
in the thesis were relevant to the happiness report and the Global key indexes including
the social, economic and educational factors. The data for 169 countries was
considered.
The thesis follows a systematic approach and including the steps of the initial data
analysis, followed by the data preprocessing and further applying the models.
Machine learning models including the regression models (Multiple, Gradient
Boosting, SVM), classification models including the random forest and KNN has been
utilized. Along with this a basic Deep learning model of ANN has been utilized to make
the approach
Key visualizations and metrics which were relevant to the output of the model were
utilized to compare the techniques. Based on the results from the approaches applied,
classification model of the KNN tends to give the most accurate results with the highest
model accuracy and better predictions.