Regression: The act of returning to an original position or state after going around in a circle.

In computer science, the term "regression" is typically used in two different contexts:

1. Machine Learning
2. Software Testing

Regression in the Context of Machine Learning

• In this context, the original meaning of the word "regression" extends to approximating relationships and making predictions.
• Here, regression primarily refers to a method of mathematically approximating relationships between variables and predicting outcome values.
• It's used as a term to describe analytical techniques that mathematically model the relationship between independent and dependent variables to predict or explain the values of dependent variables.
• For example, in predicting house prices, when you create a model that predicts the outcome variable (e.g., price) based on input variables (e.g., geographical factors, area), this can be described as constructing a regression model.

There's a historical context to how regression came to have a meaning close to prediction. One of the pioneers of modern statistics, Francis Galton, discovered that "when parents have extreme heights, their children's heights tend to regress toward the mean compared to their parents' heights" (regression toward the mean). Subsequently, people used the term regression not only to describe the phenomenon of returning to the average but also to refer to "techniques that mathematically model relationships between variables to predict values."

Regression in the Context of Software Testing

• In this context, "regression" is used in its original meaning of "returning" or "going back."
• It refers to testing that verifies whether software has returned to a state where previous bugs reappear.
• When development should be moving forward, but after modifications, old bugs reappear, we can say that the software has "regressed."