Breast cancer is a common malignant tumor of which pathogenic genes are widely studied. Since gene pairs are considered as biomarkers to identify cancer patients, in this paper, we use information theory to study the collaboration features of gene pairs. The measure of synergy based on mutual information (MI) is introduced to determine whether genes collaborate with each other in breast cancer. Part mutual information (PMI) is introduced to further select collaborative genes and construct a synergy network, which overcomes the shortage of MI. Furthermore, a dual network of synergy network is constructed and structural indices are calculated to identify vital genes. By decision tree and support vector machine, synergy is considered as a suitable index and dual network with PMI improves the accuracy of cancer identification. This method can be extended to identify other biological phenomenon and find collaborative genes as biomarkers.
A nonlinear support vector machine‐based feature selection approach for fault detection and diagnosis: Application to the Tennessee Eastman process
