摘要
基于功能域组分和分组重量编码特征提取预测参数,采用支持向量机的识别算法,在Jackknife检验下,蛋白酶的预测结果为89.9%,蛋白酶的6个类别的总预测结果为94.0%。预测结果表明,该方法是预测蛋白酶及其类别的有效工具。
By using the functional domain and the encoding based on grouped weihth,a new feature parameter is extracted for protein sequence.It is then applied to the classification of protease types along with use of the support vector machine.The overall result obtained from the jackknife test in identifying protase and non-protease reaches 89.9% and that in identifying the protease types reaches 94.0%.This method is proved to be a useful tool for predicting protease types.
出处
《内蒙古工业大学学报（自然科学版）》
2011年第2期116-122,共7页
Journal of Inner Mongolia University of Technology：Natural Science Edition
基金
内蒙古自然科学基金资助项目(2009MS0111)
国家自然科学基金项目(30960090)
关键词
蛋白酶
功能域
分组重量编码
支持向量机
protease
functional domain
encoding based on grouped weight
support vector machine
