7
fall categories and zero for all other categories which is not
accepted. Therefore, text classification using the supervised
machine learning model is a promising technique in ana-
lyzing large data in mining application. The authors consid-
ered only one predictor to classify the ground-fall incidents,
better results could be obtained if more predictors were
considered in the analysis. Additionally, the ground-fall
narratives were extracted from incident reports and some
words were misspelled which could affect the performance
of the machine learning models, future work would explore
the effect of misspelled words on the model performance
and would also consider more than one predictor in the
Figure 5. F1-score for various machine learning models to classify ground-fall narratives
Table 4. Model performance for three different machine learning algorithms
Multinomial Naïve Bayes Random forest Logistic regression
Confusion matrix
F1_Score for 0 0.42 0.62 0.72
F1_Score for 1 0.46 0.44 0.78
F1_Score for 2 0.52 0.56 0.73
F1_Score for 3 0.40 0.74 0.85
F1_Score for 4 0.84 0.95 0.98
Overall F1_Score 0.74 0.91 0.96
fall categories and zero for all other categories which is not
accepted. Therefore, text classification using the supervised
machine learning model is a promising technique in ana-
lyzing large data in mining application. The authors consid-
ered only one predictor to classify the ground-fall incidents,
better results could be obtained if more predictors were
considered in the analysis. Additionally, the ground-fall
narratives were extracted from incident reports and some
words were misspelled which could affect the performance
of the machine learning models, future work would explore
the effect of misspelled words on the model performance
and would also consider more than one predictor in the
Figure 5. F1-score for various machine learning models to classify ground-fall narratives
Table 4. Model performance for three different machine learning algorithms
Multinomial Naïve Bayes Random forest Logistic regression
Confusion matrix
F1_Score for 0 0.42 0.62 0.72
F1_Score for 1 0.46 0.44 0.78
F1_Score for 2 0.52 0.56 0.73
F1_Score for 3 0.40 0.74 0.85
F1_Score for 4 0.84 0.95 0.98
Overall F1_Score 0.74 0.91 0.96