基于条件随机场进行蛋白质二级结构预测

Protein secondary structure predict based on conditional random fields

使用一种新的概率图模型——条件随机场对蛋白质二级结构进行预测,并给出了模型的构建、训练以及解码的算法。应用这一模型对一个典型的蛋白质数据集CB513的二级结构进行了预测,并将预测结果与其他方法进行比较,预测准确度有明显的提高。

Conditional random fields(CRFs) is a new probabilistic graph model and this work introduce conditional random fields in protein secondary structure prediction.This paper gave the method of constructing the model and the algorithms to train and decode the model and used the model to predict the second structure of a famous protein dataset(CB513).Finally compared the results with some other methods.