You are working with the text-only light edition of "H.Lohninger: Teach/Me Data Analysis, Springer-Verlag, Berlin-New York-Tokyo, 1999. ISBN 3-540-14743-8". Click here for further information.

Table of Contents Appendix References Exercises
Index
See also: interactive examples, data sets, Exercise - Detection of mixtures of two different wines by PCA, Exercise - Classification of unknown wine samples by PCA 

Exercises

The following list contains all exercises which are available within Teach/Me Data Analysis. All exercises are based on data supplied with Teach/Me. Most of them can be carried out by using the built-in data laboratory.
 

Topic Reference Remarks
Artificial neural network NOx Exhausts Modeling the NOx exhaust of an experimental motor
Autocorrelation Weather Station Data Determining the time shift between the sun and the warming up of air and sea water
Chance Correlation Artifical Data See the effect of correlation by chance when the number of observations is too low.
Characterizing Distributions Artificial Data Design a data set showing a bimodal probability density function, and calculate the most important measures of location and variation.
Cluster Analysis Similar Mineral Waters Try to find the two most similar mineral waters
Correlation Artificial Correlated Data Design an artificial data set with predefined correlations
Distribution Delayed Trains What is the probability that a given measurement falls within a given range 
Distribution Alcohol Content What is the probability that a given measurement falls within a given range
LDA Counterfeit Money Use linear discriminant analysis to discriminate between genuine and counterfeit banknotes.
Missing Values Mineral Waters Estimate missing values for moderately correlated data by multiple regression
MLR Polynomial Fit Calculate a polynomial fit by means of MLR
MLR Collinearity The effect of collinear variables on MLR models
MLR Boiling Points Estimation of boiling points from chemical structures
Multivariate Modeling Henry Constants Use several multivariate modeling methods to estimate log(H) from structural descriptors.
Normal Distribution Artificial Data Design a data set showing a normal probability density function, shift and scale it, and calculate the most important descriptive parameters of it.
Outliers Creating outliers Create two different data sets and experiment with outliers
PCA Wine Blending Detecting mixtures in PCA plots
PCA Artificial Data Dependence of principal component score on scaling of data
PCA Classification of Wine Classify two unknown samples of wine
PCR Octane Numbers Apply principal component regression as a combination of PCA and MLR
Simple Regression Weight of Perch Try to Estimate the weight of perch from their body length
Simple regression Solid Residues Establish the relationship between bicarbonate concentration and solid residues of mineral waters.
Simple regression Intestine Cancer Try to set up a model for the relationship between male and female cancer data. 
t-Test Humidity of US cities Compare the relative humidities of 264 US cities 
t-Test Coins Use t-Test to decide whether coins lose weight during usage over the years
t-Test Reaction Times Test whether the reaction time of a person is above a certain threshold.
t-Test Strontium concentration in  drinking water Compare the strontium concentration of several water wells in different areas.
Variability Reaction Times Measure your reaction time and calculate mean and standard deviation

Last Update: 2005-Jän-25