Version 1.0

Release Features:

• Extreme Learning Machine implementation
• Save and Load the model
• Different activation function
• Several Metrics for binary or multiclass classification

Roadmap for the future reales:

• Improvement predictionClass ( actually is more then O(n) due to the iteration over the prediction matrix and the argmax function on the matrix row )
• Improvement of computeConfusionMatrixValues method
• Replace the usage of dense vector/matrix with their sparse variant