|
[Note: thIs transcrIptIon was produced by an automatIc OCR engIne]
94
AppendIx
a StatIstIcal tools used
AflIne adjustment of a bI-varIed protocol wIth calculatIon of the lInear regressIon and the
coeffIcIent of correlatIon.
LInear regressIon of y Into x:
Y___a.X+b=,
a = X — rnxY — my = CavarIanceX,Y
“X _ mxflz VarIam:eX
b = my —a -mx
QualIty of the adjustment
rX,Y = cosX —mxY — my
CavX, Y
Z JVarx> -JVarY>
rX, Y = 2 1 perfect adjustment
rX, Y = 0 bad adjustment
The general coeficIents at correlatIon are as follows:
KarImarra 0.09 bad generalIsed category
Purungu 0.32 bad generalIsed category
Panaka 0.44 bad generalIsed category
PaljerI — 0.66 average generalIsed category but to a lesser degree than prevIous cats.
Burgulu - 0.71 relatIvely good dIslocated category
Tjarurru 0.03 bad localIsed category
MIlangka 0.88 relatIvely good localIsed category
YIparrka0.02 bad localIsed category
|
| |