Modeling techniques

Prediction, classification (using neural networks, Bayesian networks, trees, …), regression

Clustering, segmentation

Fitting to an a priori model

Factor analysis, principal components analysis

---

Transcript

There are various modeling techniques that we can apply, things like Bayesian networks or trees, neural networks, regression analysis, etc.  We can do clustering of the data. We can think about the data in a high dimensional space, and we can look at the points the cluster towards each other, then we can think about segmenting that space into different pieces. We can think about fitting data to an a priori model.  We say, "Hmm!  The model should be roughly like this, now what are the coefficients". We can also think about using techniques like factor analysis or principal component analysis where initially we don't know what are the components - so we need to figure out which is the most important component, which is the second most important component, etc.