We can reject the hypothesis that the age does not determine whether the credit is good or not at 95% confidence. # Residual deviance: 1188.0 on 996 degrees of freedom # Null deviance: 1221.7 on 999 degrees of freedom # (Dispersion parameter for binomial family taken to be 1) Let’s try to fit the model with an interaction and without an interaction, and decide whether an interaction should be included. On the other hand, for essential things like education and repairs, we don’t expect to see that kind of difference between old and young people. It is possible that loans are more likely to be bad if a young person is taking out a loan for something like a car or a TV, because the young person may be irresponsible and buy an expensive car/TV without knowing how they’ll repay the loan. In the analysis, we will also include the purpose of the loan. Main hypothesis: the age of the customer influences the probability of the loan’s being good, controlling for the purpose of the loan (defined as essential or non-essential, with domestic appliances, repairs, education, and retraining considered essential) and the loan amount. Now, let’s set things up so that Good Credit is 1, and Bad Credit is 0. # 1 yes, registered under the customers name yes 1 # OtherDebtors ResidenceSince Property Age OtherInstallPlans Housing

< 4 years 2 female - divorced/separated/married Let’s read in the data and rename the columns and values to something more readable data (note: you didn’t have to rename the values.) german = 200 DM / salary assignments for at least 1 yearĪ30 : no credits taken/ all credits paid back dulyĪ31 : all credits at this bank paid back dulyĪ32 : existing credits paid back duly till nowĪ34 : critical account/ other credits existing (not at this bank)Ī92 : female - divorced/separated/marriedĪ122 : building society savings agreement/ life insuranceĪ123 : car or other, not in attribute AccountĪ171 : unemployed/ unskilled - non-residentĪ174 : management/ self-employed/highly qualified employee/ officerĪ192 : yes, registered under the customers name

German credit data set arff rescue