K-Means Node Expert Options

For those with detailed knowledge of k-means clustering, expert options allow you to fine-tune the training process. To access expert options, set the Mode to Expert on the Expert tab.

Stop on. Specify the stopping criterion to be used in training the model. The Default stopping criterion is 20 iterations or change < 0.000001, whichever occurs first. Select Custom to specify your own stopping criteria.

• Maximum Iterations. This option allows you to stop model training after the number of iterations specified.
• Change tolerance. This option allows you to stop model training when the largest change in cluster centers for an iteration is less than the level specified.

Encoding value for sets. Specify a value between 0 and 1.0 to use for recoding set fields as groups of numeric fields. The default value is the square root of 0.5 (approximately 0.707107), which provides the proper weighting for recoded flag fields. Values closer to 1.0 will weight set fields more heavily than numeric fields.