2018/04/15 by Niharika Gauraha, Gauraha, Niharika, Ola Spjuth +1 · 1 citation
Computer Science · Mathematics · #Machine Learning and Data Classification #Neural Networks and Applications #Statistical Methods and Inference
paper · pdf · doi:10.48550/arxiv.1804.05494
The conformalClassification package implements Transductive Conformal\nPrediction (TCP) and Inductive Conformal Prediction (ICP) for classification\nproblems. Conformal Prediction (CP) is a framework that complements the\npredictions of machine learning algorithms with reliable measures of\nconfidence. TCP gives results with higher validity than ICP, however ICP is\ncomputationally faster than TCP. The package conformalClassification is built\nupon the random forest method, where votes of the random forest for each class\nare considered as the conformity scores for each data point. Although the main\naim of the conformalClassification package is to generate CP errors (p-values)\nfor classification problems, the package also implements various diagnostic\nmeasures such as deviation from validity, error rate, efficiency, observed\nfuzziness and calibration plots. In future releases, we plan to extend the\npackage to use other machine learning algorithms, (e.g. support vector\nmachines) for model fitting.\n