Sie sind auf Seite 1von 4

Univariate

Data
Types
Categorical/numerical:
discrete/continuous
Organising Data
Frequency tables
stemplots
Graphs
Bar/Column/Histograms
% Segmented Bar Graphs
Dot Plots
Describing Data
Symmetric/skewed
Outliers
Centre
Spread
Summarizing
Numerical Data
mean , med, mode,
Q1, Q3, SD, Range,
IQR, 68% and 95% and
99.7% bell curve,
z-scores.
Boxplots
5 value summaries
Outlier Test

Bivariate
Data
Dependent or
Independent
Variables
Back to Back
Stemplots
Parallel Boxplots
And interpreting off
them. Use terms like
50 % of data...
Two Way frequency
Table and
Segmented
Barchart , ususally
%
Scatterplots
Must know Dependent and
Independent Variables
Descriptions of Scatterplots
Direction (+ - random), form
(linear), strength (strong), outliers.
Pearsons Correlation Coefficient (r)
Remember, correlation doesnt
mean causation.
Coefficient of Determination (r
2
)
Quoted as% : ...variation in
dependent variable can be
explained by variation in
independent...
Linear
Regression
Why make a model?
In order to interpolate
and/or extrapolate
Line of fit by eye
M = (y
2
y
1
)/(x
2
x
1
)
Y y
1
= m(x x
1
)
3-Median Regression Line
Finding it graphically
Finding it with formulae
Least Squares Regression Line
Arithmetic Approach using
Y=mx+c and m = rs
y
/s
x
and

Using Caluculator and y = ax + b
or y = a + bx and rewritten to use
dependent and independent
variables.

Interpret Slope and y- intercept from equation and
describe it.
Residuals
How to calculate by hand
and on graphic calculator.
Drawing the residual
Graph.
Interpreting residuals :
random or pattern implies?
Transforming to Linearity
How to make a choice of a transformation.
How to enter into calculator.
How to test fit: residuals and r
2
.
x m y c
Time Series
Types of Trends
Secular
Seasonal
Cyclic
Random
Random or secular? Fit a trend Line
Best fit by eye
3-median method
Least squares regression
The best is Least square Regression
line. If there is an outlier, 3-median is
preferred.
Smoothing
Why? To see if there is a secular
trend, so trend lines can be fit for
forecasting.
Moving Average Smoothing
Odd number Of points: Eg 3
point moving average
Even number of points Eg: 4
point moving average and
centred
Median Smoothing
Seasonal Adjustment
If data is seasonal, we seasonally adjust it to
smooth and create a more secular trend.
Be able to do both the example in our notes
and the method on p 191.
What does a seasonal index mean? i.e. 1.3
means? 0.7 means?
Sum of indices = 1
Forecasting
Deseasonalised (adjusted) =original (raw)
figure seasonal index
When you have predicted from a trend line,
you then x by seasonal index to get the
seasonal (raw) prediction.

Das könnte Ihnen auch gefallen