Kelompok : Predictator Member :
- Ajmal Kurnia-1806169433
- Muhammad Fauzi-1806280533
- Pray Somaldo-1806280571
Dataset taken from IndoSum
- F1 : sentence similarity based on unigram overlap
- F2 : paragraph similarity based on unigram overlap (unfinished)
- F3 : unique formatting (unfinished)
- F4 : cue important phrases (not used)
- F5 : sum of TF-IDF (Term Frequency - Inverse Document Frequency)
- F6 : title unigram overlap
- F7 : sentence position in the paragraph
- F8 : cue trivial phrases (not used)
- F9 : proper noun word in sentence (unfinished)
- F10 : sum of TF-ISF (Term Frequency - Inverse Sentence Frequency)
- F11 : TextRank score (unfinished)
- Baseline (lead3) => picking the first 3 sentences of the document as summary