کد مقاله | کد نشریه | سال انتشار | مقاله انگلیسی | نسخه تمام متن |
---|---|---|---|---|
505019 | 864466 | 2015 | 11 صفحه PDF | دانلود رایگان |
• Proposed method significantly improves the prediction performance for hot regions.
• We combine density-based incremental clustering with feature-based classification.
• Feature selection is used to get the best features for classification.
Discovering hot regions in protein–protein interaction is important for drug and protein design, while experimental identification of hot regions is a time-consuming and labor-intensive effort; thus, the development of predictive models can be very helpful. In hot region prediction research, some models are based on structure information, and others are based on a protein interaction network. However, the prediction accuracy of these methods can still be improved. In this paper, a new method is proposed for hot region prediction, which combines density-based incremental clustering with feature-based classification. The method uses density-based incremental clustering to obtain rough hot regions, and uses feature-based classification to remove the non-hot spot residues from the rough hot regions. Experimental results show that the proposed method significantly improves the prediction performance of hot regions.
Journal: Computers in Biology and Medicine - Volume 61, 1 June 2015, Pages 127–137