高级检索

    基于人工神经网络的汉语兼类处理方法的研究

    CHINESE SYNTACTIC CATEGORY DISAMBIGUATION WITH THE NEURAL NETWORKS

    • 摘要: 汉语兼类处理是计算机理解汉语的一个关键技术.目前广泛研究和应用的兼类处理方法大多是基于规则的.但实践表明规则处理系统在用于汉语兼类处理时效果并不尽如人意.文中首次将人工神经网络方法引入汉语兼类处理领域,选择确定了汉语兼类处理的神经网络的结构、输入信息和算法.在经过训练后,基于人工神经网络的汉语兼类处理系统在处理能力和效率上都显著超过了规则处理系统.

       

      Abstract: Chinese syntactic category disambiguation is an important and difficult aspect in Chinese understanding with computers.In Chinese syntactic category disambiguation,most of the methods being used or studied are based on rules.In practice,it is shown that the results of the Chinese syntactic category disambiguation systems based on rules are not good as the researchers expect.A new method——a method with the artificial neural networks——is first applied to Chinese syntactic category disambiguation.The information used in Chinese syntactic category disambiguation,the architecture of the artificial neural networks, and the learning algorithm of the networks are provided.Having been trained,the disambiguation system based on the artificial neural networks can process Chinese syntactic category ambiguity more effectively and rapidly than the disambiguation system based on the rules do.

       

    /

    返回文章
    返回