How to do it...

In this chapter, the Occupancy Detection dataset from the UC Irivine ML repository is used to build models on logistic regression and neural networks. It is an experimental dataset primarily used for binary classification to determine whether a room is occupied (1) or not occupied (0) based on multivariate predictors as described in the following table. The contributor of the dataset is Luis Candanedo from UMONS.

There are three datasets tobe downloaded; however, we will use datatraining.txt for training/cross validation purposes and datatest.txt for testing purposes.

The dataset has seven attributes (including response occupancy) with 20,560 instances. The following table summarizes the attribute information: