Top Page | Upper Page | Contents | About This Site | JAPANESE

Regression Analysis for Curve

Single Regression Analysis is used when there are one X and one Y. And it uses the formulation,
Y = A * X + B

It is useful. But it is the straight line. This page is made for curves.

Use Single Regression Analysis

For example, there is the data,
Y = A * X + B
It looks like a curve.

In this case, we add the variable of square of X
Y = A * X + B

Then we use the new variable as the "X",
Y = A * X + B

We use X and Y to make 2-Dimension Scatter Plot . But we use square X and Y to do regression analysis.

We can use various formulation as X, for example, square X, sin(X), exp(X), log(X) and so on.

Use Multi-Regression Analysis

General quadratic curve is
Y = A * X + B

We need the formulation of Multi-Regression Analysis . This is the point that starting data is one X and one Y but analysis data is more than two.

Multi-Regression Analysis is done using the variables, square X and X.

NEXT Prediction Interval of Regression Analysis