KORELASI DAN REGRESI

Introduction: instrumental analysis

Instrumental analysis methods – why?

• Extreme sensitivity, sometimes single molecule detection.
• Very wide concentration ranges.
• Multi-analyte capability in conjunction with suitable data handling.
• Automation: high sample throughput and low cost per sample.
• Miniaturised systems.
• Computer interfacing for control, rapid data handling, optimisation.
Calibration graphs in instrumental analysis
• The analyst takes a series of samples (preferably at least six, and
possibly several more) in which the concentration of the analyte is
known.

• These calibration standards are measured in the analytical

instrument under the same conditions as those subsequently used
for the test (i.e. the ‘unknown’) samples.

• The results are used to plot a calibration graph, which is then used
to determine the analyte concentrations in test samples by
interpolation
Calibration procedure in
instrumental analysis:

calibration points

test sample.
consider a number of aspects of plotting calibration graphs:

1. It is usually essential that the calibration standards cover the

whole range of concentrations required in the subsequent
analyses

2. It is crucially important to include the value for a ‘blank’ sample

in the calibration curve

• The calibration curve is always plotted with the instrument

signals on the vertical (y) axis and the standard concentrations
on the horizontal (x) axis.

• This is because many of the procedures to be described in the

following sections assume that all the errors are in the y-values
and that the standard concentrations (x-values) are error-free.
Where:

b is the slope of the line and α is intercept on the y-axis

The product–moment correlation coefficient

• A common method of estimating how well the experimental

points fit a straight line is to calculate the product–moment
correlation coefficient, r. simply called as the ‘correlation
coefficient’

• The value of r is given by:

• is called the covariance of the two variables x and y: it
measures their joint variation. It measures their joint variation

• If x and y are not related their covariance will be close to zero.

• The correlation coefficient r equals the covariance of x and y

divided by the product of their standard deviations, so if x and
y are not related r will also be close to zero

• r can only take values in the range

-1 ≤ r ≤ +1
• r value of -1 describes perfect negative correlation, i.e. all the
experimental points lie on a straight line of negative slope

• r value of +1 we have perfect positive correlation, all the points

lying exactly on a straight line of positive slope.

• When there is no linear correlation between x and y the value of

r is close to zero.
Example 1

Standard aqueous solutions of fluorescein are examined in a

fluorescence spectrometer, and yield the following fluorescence
intensities (in arbitrary units):
No Fluorosence Concentration (pg
intensities ml-1)
1 2.1 0
2 5.0 2
3 9.0 4
4 12.6 6
5 17.3 8
6 21.0 10
7 24.7 12

Determine the correlation coefficient, r.

Calibration plot for the data in
Misinterpretation of the correlation
coefficient, r. Broken lines are “least
Squares” straight lines
To test for a significant correlation, i.e. H0 = zero correlation,
calculate

The calculated value of t is compared with the tabulated value at

the desired significance level, using a two-sided t-test and (n – 2)
degrees of freedom.

The null hypothesis in this case is that there is no correlation

between x and y.
The line of regression of y on x

• In this section  assume that there is a linear relationship

between the analytical signal (y) and the concentration (x), and
show how to calculate the ‘best’ straight line through the
calibration graph points, each of which is subject to experimental
error

• The straight line required is calculated on this principle: as a

result it is found that the line must pass through the centroid of
the points
It can be shown that the least-squares straight line is given by:

Slope of least-squares line:

Intercept of least-squares line

The line determined is known as the line of regression of y on x

Example 2

Calculate the slope and intercept of the regression line for the data
given data
No Fluorosence Concentration (pg
intensities ml-1)
1 2.1 0
2 5.0 2
3 9.0 4
4 12.6 6
5 17.3 8
6 21.0 10
7 24.7 12
Errors in the slope and intercept of the regression
line
• The line of regression calculated in the previous section will be
used to estimate the concentrations of test samples

• The random errors in the values for the slope and intercept are
therefore important, and we need further equations to calculate
them
 Values are the points on the calculated regression line
corresponding to the individual x-values, i.e. the ‘fitted’ y-value

Standard deviation of slope

Standard deviation of intercept

The y-residuals of a regression line.
Example 3

Calculate the standard deviations and confidence limits of the slope

and intercept of the regression line in previous example
No Fluorosence Concentration (pg
intensities ml-1)
1 2.1 0
2 5.0 2
3 9.0 4
4 12.6 6
5 17.3 8
6 21.0 10
7 24.7 12
Calculation of a concentration and its random error

• Once the slope and intercept of the regression line have been
determined, it is very simple to calculate the concentration (x-
value) corresponding to any measured instrument signal (y-
value).

• But it will also be necessary to find the error associated with

this concentration estimate.

• Calculation of the x-value from the given y-value involves the

use of both the slope (b) and the intercept (a)  both these
values are subject to error

• The instrument signal derived from any test sample is also

subject to random errors.

• As a result, the determination of the overall error in the

corresponding concentration is extremely complex
• y0 is the experimental value of y from which the concentration
value x0 is to be determined,

is the estimated standard deviation of x0

• if there are m such readings, then the equation for becomes

Example 4

Using the data from Example 1, determine x0 and sx0 values and x0
confidence limits for solutions with fluorescence intensities of 2.9,
13.5 and 23.0 units (95%).

No Fluorosence Concentration (pg

intensities ml-1)
1 2.1 0
2 5.0 2
3 9.0 4
4 12.6 6
5 17.3 8
6 21.0 10
7 24.7 12
Limits of detection

The limit of detection of an analyte may be described as that

concentration which gives an instrument signal (y) significantly
different from the ‘blank’ or ‘background’ signal

blank signal, yB, plus three standard deviations of the blank, sB

Example 5

Estimate the limit of detection for the fluorescein determination

studied in the previous example

No Fluorosence Concentration (pg

intensities ml-1)
1 2.1 0
2 5.0 2
3 9.0 4
4 12.6 6
5 17.3 8
6 21.0 10
7 24.7 12