Key features overview

• The Signatures molecular descriptor [1-3] for computing descriptors from chemistry.

• Transductive Conformal Prediction (TCP) for mondrian binary classification using LibLinear and LibSVM

• Transductive Conformal Prediction (TCP) for multi-label classification using probabilistic LibSvm

• Inductive Conformal Prediction (ICP) for classification and regression.

• Aggregated Inductive Conformal Prediction (ACP) according to [4] for classification and regression using LibSVM and LibLinear

• Cross-conformal Prediction (CCP) according to [8] for regression using LibSVM and LibLinear

• Standard p-values, smoothed p-values, linear and splines interpolation of p-values.

• Cross Venn-ABERS Prediction (CVAP) according to [9] for probabilistic classification using LibSVM and LibLinear

• Significant Signature calculation and feature highlighting according to the Gradient method described in [5]

• Grid Search and cross-validation for parameter tuning, otherwise using default values according to [6]