基于民航安全信息文本挖掘的进近着陆致险因素分析

为有效识别飞行进近着陆阶段致险因素,以进近着陆不安全事件为对象,利用Python语言和文本挖掘方法对文本进行预处理,获取包含关键特征值和特征权重的高维稀疏原始特征向量集合;利用卡方统计对致险因素原始特征项进行降维,得到30个维度文本特征项,借助词云图将文本挖掘结果可视化;通过人为因素、飞机故障、环境因素和管理因素4个层...

Risk factors analysis of approach and landing based on civil aviation safety information text mining

In order to effectively identify the risk factors in the flight approach and landing stage,taking the approach and landing unsafe events as the object,the preprocessing of text was conducted by using the Python language and text mining method,and the high-dimensional sparse original feature vector set containing key feature values and feature weights was obtained.The dimension reduction of original feature items of risk factors was conducted by using the chi-square statistics,then 30 dimensional text feature items were obtained,and the word cloud graph was used to visualize the text mining results.Through the four levels of human factors,aircraft failure,environmental factors and management factors,the risk factors of approach and landing were classified,and their importance was analyzed.The results showed that the human factors accounted for a large proportion of the cause of approach and landing events or accidents.In addition,the pilot decision-making and handling skills,bad weather and airport control were the important risk factors during the approach and landing stage.The airlines and control departments should strengthen the prevention and control,thus identify and eliminate the potential accident hazards.