Three sets of experimental data to evaluate taxonomy performance

Published: 30 August 2022| Version 1 | DOI: 10.17632/pf65d6r7v6.1
Qingsong Jia


These are the three sets of experimental data used in the paper named "Method Construction and Evaluation: A More Comprehensive and Reliable taxonomy of Causes of Coal Mine Gas Explosion". "Experiment 1 Analysis results of group A.xls" and "Experiment 1 Analysis results of group B.xls" are the analysis results of the two groups of analysts in Experiment 1, which are used to calculate the consistency of the accident analysis results. "Accident Case Paragraph Code Table.xlsx" is the coding table of the original paragraph of the accident investigation report used in the second experiment. "Experiment 2_Results of the expert group and group B analysts.xlsx" contains the accident cause coding table used in Experiment 2 and the classification results of the expert group and group B analysts after coding with the coding table. This set of data is used to calculate parameters such as F1 score to assess the criterion validity of the taxonomy. "Experiment 3_Results of secondary classification of analysis results of group A by applying taxonomy.xlsx" is the data of Experiment 3, including the accident analysis results of Group A analysts in Experiment 1 and the results of secondary classification by applying the taxonomy. This set of data is used to calculate causation coverage to assess the content validity of the taxonomy.



Accident Investigation