Article directory
R2 - Methods for evaluating regression
Regression is a method of fitting a function to data. For example, we can count the number of cars in the parking lot at the gate of Wal-Mart through satellites, and we can also know Wal-Mart's sales in the corresponding period through its earnings report. Therefore, you want to establish a functional relationship between the number of cars and the quarterly earnings of Wal-Mart, so that you can speculate in stocks. However, after establishing the functional relationship between the number of cars and quarterly earnings, how should we judge the quality of the functional relationship between you and it? A commonly used parameter to measure the fitting effect is the coefficient of determination R2. This article will introduce in detail the calculation principle of R2 and the analysis of the reasons for negative values:
What is R2
R2 is used to compare the prediction error of the regression model with the simple Y=error of the mean of the sample points.
The formula for R2 is as follows:
R 2 = 1 − S