# Problem

Given data (*x*_{i}, *y*_{i}),
for *i* = 1, 2, ..., *n*,
extrapolate a value outside the range of *x* values.

# Assumptions

We will assume the data is correctly modeled by a curve on which we may either apply linear regression, or may apply a transformation to linear regression.

# Tools

We will use linear regression.

# Process

Find the least-squares regression curve, and evaluate that function at that point.

The error in our extrapolated value depends on how far
we are from the mean of the *x* values.

Copyright ©2005 by Douglas Wilhelm Harder. All rights reserved.