Lexical Tools

Optimizing 2015 SD-Rule Set - Optimum Log

I. Criteria:

  • Total valid SD-Pairs from baseline (parent only rules) is 46,950
  • Candidate child rules are:
    • Decompose occurrence rate >= 40% (used in program)
    • Candidate child rules: occurrence rate >= 25% (shown on results)
      => 35% was used first, and the optimized set are the same.
    • Candidate child rules: precision is decided by the methodology of optimization
    • If the child rule has lower precision and recall than parents, it is not a good candidate even its recall is over 25%! Such as in Cases 14 and 15, the system performance will be worse because both precision and recall are lower than parents. No need to run the program for these two cases.
  • Find the best set by comparing parent vs. child rules:
    • Only apply when child rules precision is better than parent rule
    • Higher system performance
    • If SP is the same,
      • Use parent rule to replace child rules.
      • If not parent-child rules involved, use more rules

II. Iterative Optimization Log:

IDRank: Parent-RuleRanl: Candidate Child-RulesCutoff SD-Rules Sys Accu. Rate
(Precision)
Sys Cover. Rate
(Recall)
Sys. PerfNotes
Rule No.A. RateOccr.YesNoSD-Rule
0 Parent-rule only (Baseline)No child-Rule7661.70%18811672ar$|adj|e$|noun95.19%95.71%1.9090Baseline
1.1 15 (24):
0|2072|2053|19|$|adj|ally$|adv|99.08%|100.00%
9:
1|1954|1953|1|c$|adj|cally$|adv|99.95%|94.31%
7661.70%18811672ar$|adj|e$|noun95.21%95.50%1.9071Worse
1.2 15 (24):
0|2072|2053|19|$|adj|ally$|adv|99.08%|100.00%
9:
2|1949|1948|1|ic$|adj|ically$|adv|99.95%|94.06%
7661.70%18811672ar$|adj|e$|noun95.21%95.48%1.9070Worse
2.1 25 (37):
0|2080|2036|44|$|adj|ity$|noun|97.88%|100.00%
14, 27:
1|945|938|7|c$|adj|city$|noun|99.26%|45.43% 1|728|712|16|l$|adj|lity$|noun|97.80%|35.00%
7761.70%18811672ar$|adj|e$|noun95.19%94.89%1.9008Worse
2.2 25 (37):
0|2080|2036|44|$|adj|ity$|noun|97.88%|100.00%
13:
2|944|938|6|ic$|adj|icity$|noun|99.36%|45.38%
7661.70%18811672ar$|adj|e$|noun95.15%93.37%1.8852Worse
3.1 70 (87):
0|1324|967|357|$|noun|al$|adj|73.04%|100.00%
63:
1|673|557|116|n$|noun|nal$|adj|82.76%|50.83%
7760.23%596359237al$|adj|e$|noun95.19%95.60%1.9079Worse
3.2 70 (87):
0|1324|967|357|$|noun|al$|adj|73.04%|100.00%
60:
2|621|533|88|on$|noun|onal$|adj|85.83%|46.90%
7760.23%596359237al$|adj|e$|noun95.24%95.55%1.9079Worse
3.3 70 (87):
0|1324|967|357|$|noun|al$|adj|73.04%|100.00%
59:
3|577|497|80|ion$|noun|ional$|adj|86.14%|43.58%
7760.23%596359237al$|adj|e$|noun95.26%95.47%1.9073Worse
3.4 70 (87):
0|1324|967|357|$|noun|al$|adj|73.04%|100.00%
57:
4|472|408|64|tion$|noun|tional$|adj|86.44%|35.65%
7760.23%596359237al$|adj|e$|noun95.28%95.28%1.9056Worse
4.1 40 (54):
0|572|536|36|$|verb|ion$|noun|93.71%|100.00%
33:
1|448|433|15|t$|verb|tion$|noun|96.65%|78.32%
7661.70%18811672ar$|adj|e$|noun95.22%95.49%1.9071Worse
4.2 40 (54):
0|572|536|36|$|verb|ion$|noun|93.71%|100.00%
13:
2|321|319|2|ct$|verb|ction$|noun|99.38%|56.12%
7661.70%18811672ar$|adj|e$|noun95.23%95.25%1.9048Worse
4.3 40 (54):
0|572|536|36|$|verb|ion$|noun|93.71%|100.00%
5:
3|185|185|0|ect$|verb|ection$|noun|100.00%|32.34%
7661.70%18811672ar$|adj|e$|noun95.22%94.96%1.9018Worse
5.1 57 (73):
0|263|227|36|a$|noun|an$|adj|86.31%|100.00%
No candidate child rules found! 7661.70%18811672ar$|adj|e$|noun95.19%95.71%1.9090Same
6.1 99 (118):
0|274|1|273|a$|noun|an$|noun|0.36%|100.00%
99:
1|136|1|135|ia$|noun|ian$|noun|0.74%|49.64%
7661.70%18811672ar$|adj|e$|noun95.19%95.71%1.9090Same
7.1 53 (69):
0|137|120|17|a$|noun|ar$|adj|87.59%|100.00%
46:
1|115|105|10|la$|noun|lar$|adj|91.30%|83.94%
7661.70%18811672ar$|adj|e$|noun95.20%95.68%1.9088Worse
7.2 53 (69):
0|137|120|17|a$|noun|ar$|adj|87.59%|100.00%
39:
2|69|65|4|ula$|noun|ular$|adj|94.20%|50.36%
7661.70%18811672ar$|adj|e$|noun95.21%95.59%1.9080Worse
8.1 14 (23):
0|2514|2495|19|ation$|noun|e$|verb|99.24%|100.00%
11, 2:
1|1051|1050|1|sation$|noun|se$|verb|99.90%|41.81% 1|1256|1256|0|zation$|noun|ze$|verb|100.00%|49.96%
7761.70%18811672ar$|adj|e$|noun95.20%95.31%1.9051Worse
8.2 14 (23):
0|2514|2495|19|ation$|noun|e$|verb|99.24%|100.00%
5, 2:
2|1027|1027|0|isation$|noun|ise$|verb|100.00%|40.85% 2|1249|1249|0|ization$|noun|ize$|verb|100.00%|49.68%
7761.70%18811672ar$|adj|e$|noun95.20%95.24%1.9044Worse
9.1 49 (65):
0|290|259|31|c$|adj|s$|noun|89.31%|100.00%
43:
1|281|257|24|ic$|adj|is$|noun|91.46%|96.90%
7661.70%18811672ar$|adj|e$|noun95.20%95.70%1.9091Better
9.2 49 (65):
0|290|259|31|c$|adj|s$|noun|89.31%|100.00%
39:
2|190|180|10|tic$|adj|tis$|noun|94.74%|65.52%
7661.70%18811672ar$|adj|e$|noun95.22%95.54%1.9076Worse
9.3 49 (65):
0|290|259|31|c$|adj|s$|noun|89.31%|100.00%
28:
3|172|168|4|itic$|adj|itis$|noun|97.67%|59.31%
7661.70%18811672ar$|adj|e$|noun95.23%95.51%1.9075Worse
10.1 29 (41):
0|858|837|21|ce$|noun|t$|adj|97.55%|100.00%
18:
1|847|837|10|nce$|noun|nt$|adj|98.82%|98.72%
7661.70%18811672ar$|adj|e$|noun95.22%95.70%1.9093Better
Best
10.2 29 (41):
0|858|837|21|ce$|noun|t$|adj|97.55%|100.00%
17, 21:
2|319|316|3|ance$|noun|ant$|adj|99.06%|37.18% 2|528|521|7|ence$|noun|ent$|adj|98.67%|61.54%
7761.70%18811672ar$|adj|e$|noun95.22%95.70%1.9093Same
11.1 19 (28):
0|406|401|5|cy$|noun|t$|adj|98.77%|100.00%
No candidate child rules found! 7661.70%18811672ar$|adj|e$|noun95.22%95.70%1.9093Same
12.1 20 (29):
0|2336|2307|29|e$|verb|ion$|noun|98.76%|100.00%
13:
1|2202|2193|9|te$|verb|tion$|noun|99.59%|94.26%
7661.70%18811672ar$|adj|e$|noun95.25%95.47%1.9071Worse
12.2 20 (29):
0|2336|2307|29|e$|verb|ion$|noun|98.76%|100.00%
11:
2|2099|2095|4|ate$|verb|ation$|noun|99.81%|89.85%
7661.70%18811672ar$|adj|e$|noun95.25%95.25%1.9050Worse
12.3 20 (29):
0|2336|2307|29|e$|verb|ion$|noun|98.76%|100.00%
4:
3|600|600|0|late$|verb|lation$|noun|100.00%|25.68%
7661.70%18811672ar$|adj|e$|noun95.25%95.25%1.9050Worse
13.1 9 (15):
0|1626|1625|1|ility$|noun|le$|adj|99.94%|100.00%
2:
1|1624|1624|0|bility$|noun|ble$|adj|100.00%|99.88%
7661.70%18811672ar$|adj|e$|noun95.10%92.07%1.8717Worse
13.2 9 (15):
0|1626|1625|1|ility$|noun|le$|adj|99.94%|100.00%
2:
2|1289|1289|0|ability$|noun|able$|adj|100.00%|79.27%
7661.70%18811672ar$|adj|e$|noun95.19%94.99%1.9018Worse
14.1 12 (19):
0|1011|1007|4|sis$|noun|tic$|adj|99.60%|100.00%
13, 4:
1|333|331|2|esis$|noun|etic$|adj|99.40%|32.94% 1|367|367|0|osis$|noun|otic$|adj|100.00%|36.30%
7761.70%18811672ar$|adj|e$|noun95.20%95.05%1.9024Worse