The least squares method is basically finding the line of best fit (the square of the errors form the smaller number, therefore least squares). If you have a TI graphing calculator, you do linear regression.
You can do what hallsofivy did and enter in the data using logs and still use linear...