摘要
针对航行通告中出现的Q代码和E项正文部分不规范的问题,通过自然语言处理中的文本相似度计算方法可识别出不规范航行通告。首先,基于统计机器翻译方法将航行通告E项正文部分翻译成中文并建立数据库,将Q代码翻译成中文;然后,利用Word2vec模型计算两者之间的相似度,并制定不规范航行通告识别标准。通过对收集的500条航行通告中的Q代码和E项正文进行相似度计算,设定0.7作为不规范航行通告的识别标准,经数据测试可得不规范航行通告识别准确率为96.2%,验证了基于自然语言处理的不规范航行通告识别方法的可行性。
Aiming at the non-normative problems of item Q and item E in the notice to airmen(NOTAM),a method of text similarity calculation in natural language processing is proposed to identify the non-normative NOTAMs.Firstly,translating the item E into Chinese by statistical machine translation method,and translating the item Q into Chinese by establishing a database.Then,Word2vec model is used to calculate the similarity between the two sentences,and the recognition standard of non-normative NOTAMs is established.Based on the similarity calculation of item Q and item E in 500 collected NOTAMs,0.7 is set as the benchmark of non-normative NOTAMs.The data test shows that the recognition accuracy of NOTAMs is 96.2%,which verifies the feasibility of non-normative NOTAMs recognition method based on NLP.
作者
项恒
张驰
李猛
XIANG Heng;ZHANG Chi;LI Meng(College of Air Traffic Management,CAUC,Tianjin300300,China)
出处
《中国民航大学学报》
CAS
2022年第2期14-18,共5页
Journal of Civil Aviation University of China
基金
国家自然科学基金项目(U1633109)