Can some explain regression?

Question

I need a better understand.

Answer 1

best to look it up in a statistics book.

Linear regression finds a and b in the equation expressing the relation of y as a function of x as:

y = a + bx

The term a is the intercept and b is the slope. Terms a and b are found by the method os least squares (least squared deviations of all points (x, y) from the line.

This is univariate regression (one X). Multivariate regression expresses y as a function of x1, x2 and so on.

Answer 2

If you do an xy scatterplot, then draw a line more or less through the middle of the data, there will be a difference between the y value on the line for a particular x value and the y value actually paired with that x value. Regression tries to minimize those differences. Since some of the differences have y-data > y-line and some have y-data < y-line, if you just added, the pluses and minuses would tend to cancel out. So we minimize (y-data - y-line)².

Answer 3

A regression line is the best fit straight line that takes into account all the data points. Basically it gives you a rough idea of the slope of the data points.