Identifying Incident Causal Factors to Improve Aviation Transportation Safety: Proposing a Deep Learning Approach

Table 10

An example of how multiple labels are prepared for each incident report using one-hot encoding.

HF	AC	PR	WE	CP	AP	Truth label
✓	—	—	—	✓	—

1 indicates that a factor is present, and 0 indicates that a factor is absent. Matches and mismatches of multiple labels prepared in this way can be conveniently evaluated by the Python scikit-learn library [49].