HP HP-11C - Page 67

Section 3: Numeric Functions 65
y
5 |
(6.8, )
4 4
3T (4.5.3.0)
(4.0, y)
2 1
(3.0, 2.0)
14+
(1.5, 1.0)
} i } % 5 } } X
1 2 3 4 5 6 7
Correlation Coefficient. Both linear regression and linear
estimation presume that the relationship between the x and y data
values can be approximated, to some degree, by a linear function
(that is, a straight line). The correlation coefficient (r) is a
determination of how closely your data ‘“fits” a straight line. The
correlation coefficient can range fromr = +1tor =—-1. Atr = +1
the data falls exactly onto a straight line with positive slope. At r =
—1, the data falls exactly on a straight line with negative slope. At
r = 0, the data cannot be approximated at all by a straight line.
With statistics accumulated in registers R, through Rj, the
correlation coefficient r is calculated by pressing (5.r]. The
number that appears in the displayed X-register will be a § value
(meaningless, unless you keyed in a specific x-value, as described
above). To view the correlation coefficient value (r), exchange the
contents of the X- and Y-registers by pressing [x=].
Recall from the discussion of LAST X in section 2 that (3], [s], and
do not place a copy of the x-value in the LAST X register.
However, because y is calculated from the value in the displayed
X-register, when you press [f][j.r], a copy of the x-value is placed in
the LAST X register and the stack—in all cases—Ilifts only once.