click below
click below
Normal Size Small Size show me how
STTN Test 2
Ch 4-6
| Term | Definition |
|---|---|
| Parameter | Summarising value calculated from a whole population |
| Statistic | Summarising value from a sample |
| Measure of location/position | Potray the centre of population or sample by a single number |
| Arithmic mean | Gemiddeld x bar Continues assumption: all observations appear at class midpoint |
| Mode | value with highest incidence |
| Modal interval | interval with highest f |
| Median | Middle/center of data set if arranged from smallest to largest |
| Quartiles | Divides data in 4 parts |
| Deciles | Divides data in 10 parts |
| Percentiles | Divides data in 100 parts |
| Symmetrical distribution | Left is mirror image of right Mean = mode = median |
| Skewed right/ positive | Slopes down from left to right - slide right mean>median>mode |
| Skewed left / negative | Slides left, slopes down from right to left mode>median>mean |
| Range | difference between max and min observations |
| Interquartile range | qr = q3-q1 |
| Quartile deviation | qd=qr/2 |
| Standard deviation | gives mean distance of obs from arithmetic mean |
| Sample variance | s^2 |
| Sample standard deviation | s |
| Coefficient of variation | Standard dev as % of the mean v=s/x *100 |
| Stem and leaf plot | Summarised image of data Stem: first number Leaf: last numbers |
| Five number summary | Min; Q1; Median; Q3; Max |
| Box plots | Based on 5-number summary |
| Turkey's rule | Outlier is a value more than 1.5 times the interquartile range q1-1.5qr OR q3+1.5qr |
| Relative frequencies | r=f/n |
| Scatterplot | values of independent variables are plotted against dependent variables |
| Correlation | Linear relationship between 2 variables |
| Pearsons Correlation coefficient | measures degree of linear relationship between two variables r=sxy/sxsy |
| Properties of Correlation coefficient | x-1<=r<=1 r>0: positive relationship r<0: negative relationship r=1: PERFECT positive r=-1: PERFECT negative r=0: NO linear realtionship |
| Regression | Determine relationship between x and y & to predict values |
| Regression eq | y=alpha+betha(x)+e |
| Estimated regression ew | y=a+bx |
| Method of least squares | "Best" regression line to fit data Vertical distances between data points are as small as possible |
| Residual | Diff between observed value and value given by estimated regression line |
| Interpolation vs Extrapolation | Inter: estimate y for x INSIDE interval Extra: estimate y for x OUTSIDE interval |
| Coefficient of determination | R^2 indication of how well least-squares curve fits observed data Linear: r^2 = R^2 |
| Properties of R^2 | 0<=R^2<=1 R^2 = 1: perfect fit R^2 = 0: doesn't fit at all straight line: R^2 = r^2 |