In your science experimental work, you may often wish to identify the trend describing your data, by fitting a best-fit curve to your data. Often, graphing software will create this curve for you -- also known as the multivariate regression equation -- as well as the corresponding prediction interval, which shows how well your data is described by your regression equation. However, if you wish to calculate the prediction interval, known also as the R^2 value, you may do so by hand. The calculation, although tedious, is not difficult.
Instructions
1. Calculate the distance between each y-value predicted by your best-fit curve and each y-value. Square each distance, then sum the squares.
2. Calculate the distance between each actual y-value and the mean of all of the actual y-values. Square each distance, then sum the squares.
3. Divide your number in Step 1 by your number in Step 2.
4. Subtract your number in Step 3 from one. You will obtain a decimal number, such as .73. You may multiply this decimal by 100 percent, to show that the prediction interval is 73 percent, meaning that 73 percent of your data can be described well by your regression equation.
Tags: your data, number Step, regression equation, your number, your number Step
