From ec93bf37e79ce70bfdbe8d6e5c6de457613cc84e Mon Sep 17 00:00:00 2001 From: Chris Dyer Date: Mon, 5 Nov 2012 23:41:55 -0500 Subject: larger training data for semi-crf word segmenter --- compound-split/de/weights.trained | 40 +++++++++++++++++++-------------------- 1 file changed, 20 insertions(+), 20 deletions(-) (limited to 'compound-split/de/weights.trained') diff --git a/compound-split/de/weights.trained b/compound-split/de/weights.trained index 4ae8a8ce..313d6ba9 100644 --- a/compound-split/de/weights.trained +++ b/compound-split/de/weights.trained @@ -1,20 +1,20 @@ -# Objective = 141.249 (eval count=281) -LettersSq -0.04232699523807458 -LettersSqrt 0.4355587430228624 -RevCharLM 0.41198831478844122 -FugS 0.075512682701211239 -FugN -0.61902217202456356 -WordCount -0.0082286209848003913 -InDict -0.98529136326577915 -InDictSubWord 1.0386001157542868 -Short 0.70242841302446457 -Long -0.69651861257390713 -OOV 0.97706274228074586 -OOVSubWord -0.76138571782502074 -ShortRange -1.1864424374105051 -HighFreq -4.1150415279961052 -MedFreq 0.014790338975451987 -Freq -0.28901069668114737 -Bad -3.8059407890457644 -FreqLen1 -0.3827361966178347 -FreqLen2 -0.17308899259418953 +# Objective = 164.304 (eval count=118) +WordCount 0.10973670965860199 +LettersSq -0.040894083779153829 +LettersSqrt 0.33757888570124861 +InDict 0.049573722026753933 +InDictSubWord 0.69975734577036308 +Short 0.33555471096277373 +Long -0.41832772458034762 +OOV 0.060162987632197122 +OOVSubWord 0.068471660184093452 +ShortRange -0.95127190946200446 +HighFreq -2.4179829648207223 +MedFreq -0.21881857269052826 +Freq -0.32292456278379877 +FreqLen1 -0.34734160349477439 +FreqLen2 -0.1574854779610021 +Bad -1.7868318814690027 +RevCharLM 0.52613329631178929 +FugS 0.028876419364008043 +FugN -0.086060954349416352 -- cgit v1.2.3