R-SQUARED: Equivalent Test-Statistic Where The Mean is a Moving Average as Opposed to Linear Regression Line?

Question

If R-squared measures the relationship between your dependent Y value and independent X value for a Linear Regression Model, is there an equivalent test-statistic when you use a Simple Moving Average Model?

Answer 1

You can use MSE to do a similar test. For moving average you would take x-hat, the value predicted by your model and subtract x, you're actual value. square the difference, divide by x-hat, sum these and then use the chi-squared test statistic.

Answer 2

Averaging shouldn't provide comparable R sq. is right. it additionally will selection in case you elementary 4 values and use the averages particularly than 3 values. it fairly is, what proportion values you used for each elementary might additionally effect the fee of R^2.