Polls Regression

Data regression is an important task that can be applied to continuous data, in search for the best modeling.

​​We apply here smoothing techniques, local weighted regression and regression tree to fit the data. The goal here is just to show different approaches that can be applied.

​The data used for this analysis is the 2008 U.S. presidential elections and we consider the presence of only uncertainty in the data. A complete Bayesian statistics analysis was performed in the previous section.