Linear Regression
Introduction
Linear regression is one of the most widely used methods originating from statistics. It is particularly valuable in situations where the goal is to predict a numerical output variable, such as forecasting stock prices, temperatures, or sales. As a result, linear regression finds applications in diverse fields including economics, climate science, and marketing.
This chapter focuses on linear regression, a fundamental approach to supervised learning. Specifically, it provides a method for predicting a quantitative response. Despite its long history and the perception that it may be less exciting than some modern statistical learning techniques, linear regression remains a widely used and practical tool.
Importantly, understanding linear regression lays the foundation for more advanced methods. Many contemporary statistical learning approaches can be viewed as generalizations or extensions of linear regression. Therefore, a solid grasp of the concepts and techniques underlying linear regression, particularly the least squares approach commonly used for fitting these models, is essential before moving on to more complex learning methods.
Suppose that in our role as statistical consultants, we are asked to suggest, based on available data, a plan for improving outcomes in the social sector for the coming year. What information would be useful in order to provide such recommendations? Here are a few important questions we might seek to address:
- Education
Social Sector Expenditure and Human Development in Indian States This study utilized panel regression analysis to assess the impact of per capita social sector expenditure on the Human Development Index (HDI) across 16 Indian states from 1990 to 2018. The findings indicated that investments in education and health significantly contribute to human development, with social sector expenditure proving more effective than per capita income in enhancing HDI.(Agarwal, 2015)
- Healthcare
Simple Linear Regression
Hrllp