GNU/Linux Desktop Survival Guide
by Graham Williams
MLHub Rain Tomorrow
Decision trees and the ensemble of decision trees within a random forest are two common approaches to building classification models in AI. The concept of an ensemble of decision trees was introduced in 1988 in the paper Combining Decision Trees: Initial results from the MIL algorithm where the improved performance from multiple trees is demonstrated. The rattle package in R provides the weatherAUS dataset which is used to predict rain tomorrow. This is the dataset used to build the model that is demonstrated in these MLHub packages: rain and rainrf.
These MLHub packages also demonstrate how to deliver multiple MLHub models from a single git repository. This one repository contains a yaml file for the rain model (MLHUB.yaml) and another yaml file for the rainrf model (rainrf.yaml).
Here we install, configure and demonstrate the rain model, a decision tree model for predicting whether it will rain tomorrow:
$ ml install rain $ ml configure rain $ ml demo rain