Home arrow Paper results

Papers download
  • "A Word and Character-Cluster Hybrid Model for  Thai Word Segmentation", Canasai Kruengkrai, Kiyotaka Uchimoto, Jun’ichi Kazama,Kentaro Torisawa, Hitoshi Isahara, and Chuleerat Jaruskulchai. >>download<<
  • "Thai Word Segmentation using Character-level Information", Kessaraporn Suesatpanit, Proadpran Punyabukkana, and Atiwong Suchato. >>download<<
  • "TLex: Thai Lexeme Analyser Based on the Conditional Random Fields", Choochart Haruechaiyasak and Sarawoot Kongyoung. >>download<<
  • "A Statistical-Machine-Translation Approach to Word Boundary Identification: A Projective Analogy of Bilingual Translation", Phiradet Bangcharoensap, Peerachet Porkaew, and Thepchai Supnithi. >>download<<
  • "Thai Word Segmentation based-on GLR Parsing Technique and Word N-gram Model", Piya Limcharoen, Cholwich Nattee and Thanaruk Theeramunkong.  >>download<<
*InterBEST 2009: Thai Word Segmentation Workshop
Proceedings of 2009 Eighth International Symposium on Natural Language Processing (SNLP2009), October 20-21, 2009 Bangkok Thailand.
 
12 Genre Evaluation Results Table (Precision, Recall and F-measure)
12 Measure Limcharoen Bangcharoensap Haruechaiyasak Suesatpanit Kruengkrai
Genre

Precision 88.92 93.12 95.71 96.20 98.58
Article Recall 94.44 97.24 96.54 97.26 97.10

F-measure 91.60 95.14 96.13 96.73 97.84

Precision 90.84 96.19 95.19 97.20 98.85
Buddhism Recall 96.20 98.09 94.96 97.63 98.20

F-measure 93.44 97.13 95.07 97.42 98.52

Precision 90.11 93.54 95.15 96.37 98.14
Encyclopedia Recall 94.72 96.52 94.83 96.60 97.26

F-measure 92.36 95.01 94.99 96.48 97.70

Precision 82.35 94.33 95.36 97.83 98.93
Law Recall 91.05 97.77 96.00 98.40 97.53

F-measure 86.48 96.02 95.68 98.11 98.23

Precision 77.43 87.42 93.54 94.95 97.78
News Recall 92.61 94.84 94.99 96.13 96.30

F-measure 84.34 90.98 94.26 95.54 97.03

Precision 87.03 92.64 95.43 94.68 97.37
Novel Recall 93.99 96.31 95.72 95.46 96.20

F-measure 90.38 94.44 95.57 95.07 96.78

Precision 91.66 95.48 96.99 97.57 98.34
Talk Recall 95.99 97.63 96.78 97.39 98.20

F-measure 93.78 96.55 96.89 97.48 98.27

Precision 77.35 86.47 91.62 94.18 97.26
Wiki Recall 90.31 94.64 93.58 95.07 95.38

F-measure 83.33 90.37 92.59 94.62 96.31

Precision 89.85 93.54 95.83 96.29 97.80
NSC Recall 93.53 93.60 95.53 96.01 97.94

F-measure 91.65 93.57 95.68 96.15 97.87

Precision 74.39 85.86 92.09 91.38 95.60
Old document Recall 88.44 93.51 92.81 91.51 94.22

F-measure 80.81 89.52 92.45 91.44 94.91

Precision 60.12 69.71 84.41 89.82 96.08
Royal news Recall 87.46 91.91 92.28 93.76 88.81

F-measure 71.26 79.28 88.17 91.75 92.30

Precision 78.47 86.43 92.29 95.17 97.60
TV news Recall 92.91 95.44 94.92 96.04 95.54

F-measure 85.08 90.71 93.59 95.60 96.56

Precision 82.38 89.56 93.63 95.14 97.69
Average Recall 92.64 95.63 94.91 95.94 96.06

F-measure 87.04 92.39 94.27 95.53 96.86