1. Build a "dataset" from different datastorage, e.g., relational model or text on the Web, considering its structure and semantics in order to draw hypotheses and interpret results;
2. Prepare data via denormalization, assembling and discretization;
3. Explore the characteristics, options, benefits and limitations of supervised classification methods: a) with statistical support, b) based on the induction of decision trees, c) based on competitive learning;
4. Introduce time series analysis; adapt datastet to apply (in this context) supervised classification methods
5. Explore unsupervised methods based on instances;
6. Explore the methods that search for association rules and highlight the difference between those methods and the ones related to classification and clustering;
7. Evaluate learning via error estimation supported on the concepts of training, validation and testing sets; comparison of models and results presentation.

Curricular Unit Form