中文摘要：

针对大规模知识库问答的特点,构建了1个包含3个主要步骤的问答系统：问句中的命名实体识别、问句与属性的映射和答案选择。使用基于别名词典的排序方法进行命名实体识别,使用结合注意力机制的双向LSTM进行属性映射,最后综合前2步的结果从知识库中选择答案。该系统在NLPCC-ICCPOL 2016 KBQA任务提供的测试数据集上的平均F1值为0.809 7,接近已发表的最好水平。

英文摘要：

To solve the specific problem in KBQA,the paper builds a question answering system based on large scale Chinese knowledge base. This system consists of three main steps： recognition of named entity in question,mapping from question to property in KB,and answering selection. In the research,use alias dictionary based ranking method to recognize named entity contained in question,and attention mechanism with bidirectional LSTM for question-property mapping.Finally,exploit results of first two steps to select the answer from knowledge base. The average F1 value of this system in NLPCC-ICCPOL 2016 KBQA task is 0.809 7,which is competitive with the best result.