A linear regression channel consists of a median line with 2 parallel lines, above and below it, at the same distance. Those lines can be seen as support and resistance. The median line is calculated based on linear regression of the closing prices but the source can also be set to open, high or low. The height of the channel is based on the deviation of price to the median line. Extrapolating the channel forward can help to provide a bias and to find trading opportunities.
about Linear Regression Channels.