|
Twitter
|
|---|
|
system
|
Average accuracy
|
Average F1
|
|---|
|
Pratama and Sarno [34]
|
65.00%
|
–
|
|
Ong et al. [32]
|
74.23%
|
–
|
|
Ong et al. [31]
|
70.50%
|
–
|
|
Ergu I [14]
|
75.7%
|
–
|
|
Experiment model
|
|
Scenario 1–8
|
72.70%
|
0.701
|
|
Scenario 9–16
|
72.25%
|
0.698
|
|
Scenario 17–24
|
72.67%
|
0.701
|
|
Scenario 25–32
|
75.57%
|
0.736
|
|
Scenario 33–40
|
75.40%
|
0.735
|
|
Scenario 41–48
|
75.51%
|
0.736
|
|
Scenario 49–56
|
76.98%
|
0.757
|
- The highest values for the average accuracy and f1-measure for each personality model are shown in bold