Fig. 4
From: Machine learning predicts nucleosome binding modes of transcription factors
Impact of \(\alpha\) on the performance of proposed methods measured by various performance metrics such as a micro-AUPR, b accuracy. \(\alpha \in \left[ {0, 1} \right]\) controls the impact of similarity between mean vectors and covariance matrices. If \(\alpha = 0\), only the similarity between covariance matrices is considered. In contrast, the only similarity between mean vectors is considered when \(\alpha = 1\)