Saturday, 17 October 2009

The bands of linear regression

Indicator Strips linear regression arbitrarily smooths the price data is the regression result, if projected regression line, and then randomly creates band standard deviations above and below the regression line. First, data based on the selected price, smoothed, using the period and type of moving average. If you prefer not to use any smoothing, the chosen period 1. Then, the data obtained are used to form a regression line that ends in each bar, using the period of regression. Values in each bar may optionally be predicted value determined by the design of the regression line at X bar forward, where X is the projection period (if X = 0, then no prediction will not happen). Then, above and below the regression line can be held strip of standard deviations, based on a specified number of standard deviations and significance of standard deviation, calculated from data in the range of the period of regression.

In its most basic form, without smoothing (Moving Average of period 1) and without the forecast (the forecast period 0), the indicator provides a simple end point of linear regression line, which ensures the end of each bar with a period of regression. This alone provides a good replacement of the rolling average, and in fact identical to the type of moving average of least squares.

No comments:

Post a Comment