Using This Data and R, Your Job Is To:

• Decide on what data mining technique(s) would be appropriate in assessing whether there are groups of variables that convey the same information and how important that information is? Conduct such an analysis.

• Comment in your presentation on the distinct goals of profiling the characteristics of bankrupt firms versus simply predicting (black box style) whether a firm will go bankrupt and whether both goals, or only one, might be useful. Also comment on the classification methods that would be appropriate in each circumstance.

• Explore the data to gain a preliminary understanding of which variables might be important in distinguishing bankrupt from nonbankrupt firms. (Hint: As part of this analysis, use side-by-side boxplots, with the bankrupt/not bankrupt variable as the x variable.)

• Using your choice of classifers, use R to produce several models to predict whether or not a firm goes bankrupt, assessing model performance on a validation partition. Based on the above, comment on which variables are important in classification, and discuss their effect.