// This example script trains a Weka-based model on the 
//  formation energies of a small set of compounds

// Load in a dataset of compounds
data = new data.materials.CompositionDataset
data import ./datasets/small_set.txt

// Get only the ground-state instances of each compound
data duplicates RankingDuplicateResolver minmize PropertyRanker energy_pa SimpleEntryRanker

// Define where to find elemental property data
data attributes properties directory ./lookup-data/
// Select which set of elemental properties to use for attributes
data attributes properties add set general

// Generate new attributes
data attributes generate

// Set formation energy as property to be modeled
data target delta_e

// Create a simple model using Weka, then test it
model = new models.regression.WekaRegression trees.REPTree -L 2 
model train $data
model crossvalidate $data 10

// Print out training statistics
print model validation stats
print model model

// Print out suggested citations
citations

exit