Applying data mining technology into civil aviation events analysis can help to obtain the core contents quickly. Based on events caused by the crew members, the term frequency-inverse document frequency (TF-IDF) algorithm was used to vectorize the text features and give weights. The K-means algorithm was used to cluster analysis on the vectorized data, and the results obtained by the multidimensional scaling algorithm were visually displayed. Then, statistical analysis of the stages and causes of the events was carried out based on the cluster results. Finally, “mistakes of crew members” was taken as an example to do analysis. The results show that emphases should be paid on the three types of crew caused events, which are “approach suspension/go around”, “incomplete predetermined configuration” and “overspeed.” Furthermore, it proves that text feature extraction and cluster analysis can be applied to do analysis of human-caused events and help to conveniently and quickly clarify the direction of safety risk management.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Text Feature Extraction and Clustering Analysis of Events Caused by the Cockpit Crew


    Contributors:
    Liu, Junjie (author) / Ye, Yinghao (author) / Du, Yinlan (author)


    Publication date :

    2020-10-14


    Size :

    659383 byte




    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    Cockpit crew research

    Kanki, Barbara G. | NTRS | 1991


    Crew centered cockpit design methodology

    STOLLINGS, MICHAEL / AMELL, JOHN | AIAA | 1992


    Advanced crew station integration cockpit

    Pencikowski, P. | Tema Archive | 1990


    Representing cockpit crew decision making

    Klein, Gary A. / Thordsen, Marvin L. | NTRS | 1991