; verbose (-v: yes -v-: no) -v - ; keep intermediary files (-x: yes -x-: no) -x - ; flex rules (input file, binary format) ;-b (not specified, not doing actions with already created flex rules.) ; Word list (Optional if -b is specified. Otherwise N/A) (-I filename) ;-I (N/A) ; Output, lemmas of words in input (-I option) ;-O (N/A) ; word/lemma list -i latin.cleaned.ph ; extra file name affix -e ziggurat ; suffix only (-s: yes -s-: no) -s - ; make rules with infixes less prevalent(-A: yes -A-: no) -A - ; columns (1 or F or W=word,2 or B or L=lemma,3 or T=tags,0 or O=other) -n FBO ; max recursion depth when attempting to create candidate rule -Q 1 ; flex rules (output, binary format, can be left unspecified) ;-o (Not specified, autogenerated) ; temp dir (including separator at end!) -j tmp/ ; penalties to decide which rule survives (4 or 6 floating point numbers: R=>R;W=>R;R=>W;W=>W[;R=>N/A;W=>NA], where R=#right cases, W=#wrong cases, N/A=#not applicable cases, previous success state=>success state after rule application) -D -0.0012861551;-0.6999562655;0.7108840224;-0.0674097817;-0.0007561510;0.0126027136; ; compute parms (-p: yes -p-: no) -p ; expected optimal pruning threshold (only effective in combination with -XW) -C -1 ; tree penalty (-XC: constant -XD: more support is better -XE: higher entropy is better -XW: Fewer pattern characters other than wildcards is better) -X C ; current parameters (-P filename) -P parms.txt ; best parameters (-B filename) -B best_ziggurat.txt ; start training with minimal fraction of training pairs (-Ln: 0.0 < n <= 1.0) -L 1.000000 ; end training with maximal fraction of training pairs (-Hn: 0.0 < n <= 1.0) -H 1.000000 ; number of differently sized fractions of trainingdata (natural number) -K 20 ; number of iterations of training with same fraction of training data when fraction is minimal (positive number) -N 100.000000 ; number of iterations of training with same fraction of training data when fraction is maximal (positive number) -M 10.000000 ; competition function (deprecated) ;-f (N/A) ; redo training after homographs for next round are removed (-R: yes -R-: no) ;-R - (N/A) ; max. pruning threshold to evaluate -c 5 ; test with the training data (-T: yes -T-: no) -T ; test with data not used for training (-t: yes -t-: no) -t ; create flexrules using full training set (-F: yes -F-: no) -F ; Number of clusters found in word/lemma list: 5482 ; Number of lines found in word/lemma list: 18169 ; Evaluation: ; ----------- ; Lemmatization results for all data in the training set. ; For pruning threshold 0 there may be no errors (diff%%). ; prun. thrshld. 0 1 2 3 4 5 ; rules 7750.000000 4197.000000 974.000000 509.000000 359.000000 280.000000 ; rules% 42.655072 23.099785 5.360779 2.801475 1.975893 1.541086 ; same% 93.598987 76.878199 62.034234 58.610821 56.761517 55.435082 ; ambi1% 3.098685 3.384886 1.777753 1.106280 0.902636 0.787055 ; ambi2% 3.098685 3.038142 1.557598 0.886125 0.627442 0.566900 ; ambi3% 0.203644 0.082558 0.000000 0.000000 0.000000 0.000000 ; diff% 0.000000 16.616214 34.630414 39.396775 41.708404 43.210964 ; same%stdev 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 ; ambi1%stdev 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 ; ambi2%stdev 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 ; ambi3%stdev 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 ; diff%stdev 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 ; ;Evaluation of prediction of ambiguity (whether a word has more than one possible lemma) ;--------------------------------------------------------------------------------------- ; amb.rules% 6.401013 7.485277 5.047058 3.252793 2.603335 2.289614 ; false_amb% 0.000000 3.566514 3.704111 2.515273 2.047443 1.805273 ; false_not_amb% 0.000000 2.482250 5.058066 5.663493 5.845121 5.916671 ; true_amb% 6.401013 3.918763 1.342947 0.737520 0.555892 0.484341 ; true_not_amb% 93.598987 90.032473 89.894876 91.083714 91.551544 91.793715 ; precision 1.000000 0.354582 0.153459 0.127863 0.119527 0.118280 ; recall 1.000000 0.612210 0.209802 0.115219 0.086844 0.075666 ; Evaluation: ; ----------- ; Lemmatization results for data that is not part of the training data. ; prun. thrshld. 0 1 2 3 4 5 ; rules 7654.033333 4146.633333 958.800000 504.400000 353.600000 275.366667 ; rules% 42.758530 23.164773 5.356245 2.817783 1.975353 1.538310 ; same% 51.366120 53.278689 54.371585 54.384004 53.365623 52.769498 ; ambi1% 2.247889 1.999503 1.142573 0.956284 0.782414 0.583706 ; ambi2% 1.726279 1.577248 1.130154 0.956284 1.018381 0.782414 ; ambi3% 0.024839 0.037258 0.012419 0.012419 0.000000 0.000000 ; diff% 44.634873 43.107303 43.343269 43.691008 44.833582 45.864382 ; same%stdev 4.392414 4.819260 5.072578 5.128338 5.128445 5.126362 ; ambi1%stdev 0.854821 0.783972 0.736214 0.678932 0.712987 0.618955 ; ambi2%stdev 0.835522 0.543366 0.600594 0.487147 0.462156 0.495978 ; ambi3%stdev 0.131348 0.149053 0.065674 0.065674 0.000000 0.000000 ; diff%stdev 4.845955 5.313427 5.373637 5.434234 5.321955 5.442859 ; ;Evaluation of prediction of ambiguity (whether a word has more than one possible lemma) ;--------------------------------------------------------------------------------------- ; amb.rules% 7.935917 7.128664 4.694486 3.688525 3.440139 2.682563 ; false_amb% 0.161451 0.136612 0.099354 0.074516 0.074516 0.074516 ; false_not_amb% 0.049677 0.049677 0.049677 0.049677 0.049677 0.049677 ; true_amb% 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 ; true_not_amb% 2.521113 2.545951 2.583209 2.608048 2.608048 2.608048 ; precision 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 ; recall 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 ; ; Power law relating the number of rules in the decision tree to the number of examples in the training data ;---------------------------------------------------------------------------------------------------------- ; #rules = 1.036*N^0.909 0.563*N^0.910 0.073*N^0.973 0.032*N^0.994 0.017*N^1.020 0.007*N^1.103 ; Postscriptum ; The number of rules can be estimated from the number of training examples by ; a power law. See the last line in the table above, which is based on 7 ; different samples from the total available training data mass varying in size ; from 1.54 % to 98.56 %