mguldner / data_algorithms Goto Github PK

The Iris Dataset can be used to improve these methods.

For now, I combine the two datasets because some value for fare in test set are higher than the highest value in training set.
Solutions :

• if higher value, add a new Frame with top bound is this new value ? And if already an other value in test set that has this behavior, do not add a new Frame but just increase this one? But problem because it's also impossible to use the trained model to cluster these values...

• if higher value, put in Frame -1 <= x <= -1. But loss of interest...

This will enable to test and approve the implementation of clustering algorithms.