The glass data
Ian Witten uses Weka to browse around a more realistic dataset, interpreting the attributes, and shows what the data file itself looks like
The glass dataset is a more realistic dataset with 214 instances and 10 attributes. Each instance represents a piece of glass, and its class is the type of the glass. There are 7 possible types, corresponding to different glass manufacturing processes. We interpret the attributes, and check their values for reasonableness. Weka datasets use a format called ARFF, and we take a look at the raw glass.arff data file.
