|
As a field of data mining is an important branch , has been widely used classification technique , and after years of research and development, many classical classification methods have been researchers are familiar with, such as k-nearest neighbor , Bayesian methods , decision trees, support vector machines, neural networks. These traditional methods have some limitations , so researchers have accordingly proposed a multi- classifier systems, while multiple classifier systems research is faced with a number of important issues. Protein function prediction as the post-genomic era one of the major challenges facing many machine learning algorithms gradually been studied out. G protein-coupled receptors (G-protein coupled receptors, GPCRs) are a class of very important signal molecule receptor , because it can regulate G protein activity combined with the name. Since the structural features of GPCRs and their important role in signal transduction , determines that it can serve as drug targets , the current top-selling drugs , 20% belong to GPCRs related drugs , the world's drug market is about 1 /3 of the small-molecule drugs are of GPCRs activators or antagonists . In addition , GPCRs dysfunction can cause a variety of diseases produced . Thus, the function of GPCRs relevant data has an extremely important application value. In this paper, data mining related technologies, through the theory and practice of previous research results , for the realization of multiple classifier systems are the main research questions , put forward the corresponding improvement plans and strategies , and weka data mining based analysis platform programming the system , and the function of GPCRs for handling and processing data , experimental results show that the classification system performance has improved to some extent .
|