Laserfiche WebLink
',it.~ r::.. _•.ld fnr co\'i'1' ui•: ~ .. id na tint r:c:..., c.'.~:~c•'.":._~ ,.~ r}:.., C~ <br />i ;.~cics cnlc and \.'i]] be e=: r.i~ii:ili~d ~;~- linear r~; ,, =irn ti-c'.~nicces. iaic .,~..: <br />reeuires a rinir..um of three oars of sa:~ple data to ce\~clop a linear regression <br />model. Tile follo\:ins e>:arple based on lfendenhal] (1y75) illustrates the approach: <br />fear <br />1 2 3 4 <br />$ite X X 7( X <br />l 2 3 4 <br />1. Reference area (xi) 50 60 45 40 <br />2. Premine areas ( }•i) 45 65 45 35 <br />From the historical data, a simple linear regression equation can be written, i.e. <br />)' Bo + Blxi <br />n _ _ n _ <br />where: B1 = SS aSSx and: SSx} = E (xi - x) (yi - y); SSx = E (xi - x)Z <br />~ i=1 i=1 <br />Bo=}' - B1 x <br />The ~•ariance ma}• be estimated (s Z) by: <br />s2 = SSE = (n-2) <br />n _ <br />where: SSE =R (Yi - Y)Z - B1SSx"Y <br />i=1 <br />n = number of xi or yi observations used to develop the model, four in <br />this case. <br />Osins the formulae for the example given: <br />E _ -22.]43, Bl = 1.429, r = 0.969, sz = 14.21 <br />0 <br />y = -22.]43 + 1.429x. <br />i <br />8/12/82 <br />4-107 <br />