Table 1 Example of average set-based agreement calculated over different top-P depths (P = 1,2,3,4,5)
From: Quality indices for topic model selection and evaluation: a literature review and case study
Depth (P) | Tokens in Bootstrap Sample-S at Depth-P | Tokens in Bootstrap Sample-T at Depth-P | Intersection of Tokens in Samples S and T | Fraction Agreement (AP) |
|---|---|---|---|---|
1 | Cough | Flu | {} | 0/2 (0) |
2 | Cough, Cold | Flu, Cold | {Cold} | 1/3 (33.3) |
3 | Cough, Cold, Flu | Flu, Cold, Cough | {Cough, Cold, Flu} | 3/3 (100.0) |
4 | Cough, Cold, Flu, Fever | Flu, Cold, Cough, Phlegm | {Cough, Cold, Flu} | 3/5 (60.0) |
5 | Cough, Cold, Flu, Fever, Phlegm | Flu, Cold, Cough, Phlegm, Fever | {Cough, Cold, Flu, Fever, Phlegm} | 5/5 (100.0) |