中文摘要：

真核启动子预测是DNA序列分析的重要问题之一．本文提出一个基于内容的新算法，考虑DNA的双链特征，特征的出现频率用位置权矩阵表示，利用boosting算法进行训练和预测，得出的结果与其他算法相比，有很好的敏感性和特异性．

英文摘要：

The prediction of Eukaryotic promoter is one of the most important problems in DNA sequence analysis. In this work, we develop a new algorithm which is based on word considering the double-stranded characteristics of DNA. The frequency of occurrence of characteristics is represented by positional weight matrix. The boosting algorithm is used to train and predict respectively. The results in the algorithms have better sensitivity and specificity than that of the other algorithms.