Chapter 10. Making predictions
This chapter covers
- Classification fundamentals using decision trees, Naïve Bayes, and belief networks
- Building predictors using regression and neural networks
- Leveraging the WEKA APIs for making predictions
- Classification and regression using JDM APIs
In this chapter, we build predictive models. A predictive model makes a prediction for the value of an output attribute using the values associated with other input attributes. Predictive models can be categorized into two types based on whether the predicted attribute is continuous or discrete. When the predicted attribute is discrete, the problem is one of classification, whereas when the attribute is continuous, the problem is one of regression. Some predictive models, as in the case of neural networks, can be built to predict multiple output attributes, while others predict a single attribute.