Fig. 4From: Topology preserving stratification of tissue neoplasticity using Deep Neural Maps and microRNA signaturesThe confusion matrix of classification of samples for combined tissue-of-origin and neoplasticity status, in test data. The colour of each square is proportional to the number of samples classified as the label of that square. The diagonal of the matrix is dark, indicating a high number of correctly classified samples. Non-neoplastic classes generally have fewer misclassifications. A number of neoplastic skin tissue samples are misclassified as non-neoplastic skin tissue (green). Multiple classes are predicted to be neoplastic breast tissue (orange). Three classes are identified in red as not having any nodes assigned them. Above, + represents neoplastic tissue, and − represents non-neoplastic tissueBack to article page