Table 1.
Summary of the papers found related to temperament classification.
Table 2.
Meta-attributes used in the TECLA framework.
Fig 1.
MBTI classification scheme: four decomposing classifiers are trained.
Fig 2.
Example of the classifier representation used in TECLA for the MBTI model.
Fig 3.
Keirsey classification scheme: four binary classifiers are trained.
Table 3.
Distribution of users for each MBTI type.
Table 4.
Ratio between the various MBTI types of users.
Table 5.
Proportion of users by temperament in the dataset collected.
Table 6.
Average (mode) value for each attribute extracted by Plank.
Table 7.
Accuracy (ACC), F-measure (F) and AUC for Twitter with 5 features.
Table 8.
Accuracy (ACC), F-measure (F) and AUC for MRC with 9 features.
Table 9.
Accuracy (ACC) and F-measure (F) for LIWC with 25 features.
Table 10.
Accuracy (ACC), F-measure (F) and AUC for ONLP with 24 features.
Table 11.
Accuracy (ACC), F-measure (F) and AUC for the Random Forest.
Table 12.
Accuracy (ACC), F-measure (F) and AUC for Twitter with 5 features in the MBTI prediction.
Table 13.
Accuracy (ACC), F-measure (F) and AUC for MRC with 16 features in MBTI prediction.
Table 14.
Accuracy (ACC), F-measure (F) and AUC for LIWC with 27 features in MBTI prediction.
Table 15.
Accuracy (ACC), F-measure (F) and AUC for ONLP with 22 features in MBTI prediction.
Table 16.
Accuracy (ACC), F-measure (F) and AUC for the Random Forest in MBTI prediction.
Table 17.
Comparing with MBTI and Keirsey Results from the Literature.