What to deliver?

A report containing:

1. Raw data
   • describe variables according to their types: interval-scaled, binary, nominal, ordinal, ratio-scaled. Be aware that there are specific methods suitable to each type of variable.
   • Preliminary analysis (summaries, histograms, boxplots, spread measures, density). These are interesting to be applied to the raw data to ``uncover'' inconsistencies, outliers, duplicates etc.
   • List of main changes needed to be performed with the raw data.
2. Preprocessed data
3. Basic description (summaries, histograms, boxplots, spread measures, density).
4. Analysis
   • Bivariate analysis (correlations, regression)
   • Multivariate analysis (multiple variable regression, mutual information, cluster analysis - once more, be aware that some methods used to calculate similarity (dissimilarity) depend on the type of the variable. In the context of this dataset, when performing custer analysis we are interested to know if there are groups of similar patients)
   • Predictive Models (use of supervised machine learning. Suggestion: use WEKA)
     • Decision Tree learning
     • Support Vector Machines
   • Comparison
5. Discussion and Main Conclusions

This work is to be performed by groups of at most two people.