A hybrid strategy to protein name recognition

This paper presents a comprehensive approach to identifying protein name in biomedical texts. The new method integrated the Generalized Winnow algorithm and the heuristic rules to implement of initial detection of protein name. Moreover, the system introduced a statistic method to analyses the reliability of recognized protein boundary, which can be then used for expanding protein boundary which has low confidence. The experimental results show that this algorithm improves the whole performance for protein name recognition and that effective performance can be achieved in identifying boundary of protein name.