bishop.trigram.lm

It is the Speech recognition software. It is platform independent. To execute the source code,

FRONTMOST
#############################################################################
## Copyright (c) 1996, Carnegie Mellon University, Cambridge University,
## Ronald Rosenfeld and Philip Clarkson
#############################################################################
=============================================================================
===============  This file was produced by the CMU-Cambridge  ===============
===============     Statistical Language Modeling Toolkit     ===============
=============================================================================
This is a 3-gram language model, based on a vocabulary of 170 words,
  which begins "</s>", "<s>", "I"...
This is an OPEN-vocabulary model (type 1)
  (OOVs were mapped to UNK, which is treated as any other vocabulary word)
Good-Turing discounting was applied.
1-gram frequency of frequency : 54 
2-gram frequency of frequency : 438 110 49 30 24 12 11 
3-gram frequency of frequency : 999 212 86 45 39 22 14 
1-gram discounting ratios : 0.98 
2-gram discounting ratios : 0.40 0.60 0.78 1.00 0.52 
3-gram discounting ratios : 0.30 0.52 0.63 
This file is in the ARPA-standard format introduced by Doug Paul.

p(wd3|wd1,wd2)= if(trigram exists)           p_3(wd1,wd2,wd3)
                else if(bigram w1,w2 exists) bo_wt_2(w1,w2)*p(wd3|wd2)
                else                         p(wd3|w2)

p(wd2|wd1)= if(bigram exists) p_2(wd1,wd2)
            else              bo_wt_1(wd1)*p_1(wd2)

All probs and back-off weights (bo_wt) are given in log10 form.

Data formats:

Beginning of data mark: \data\
ngram 1=nr            # number of 1-grams
ngram 2=nr            # number of 2-grams
ngram 3=nr            # number of 3-grams

\1-grams:
p_1     wd_1 bo_wt_1
\2-grams:
p_2     wd_1 wd_2 bo_wt_2
\3-grams:
p_3     wd_1 wd_2 wd_3 

end of data mark: \end\

\data\
ngram 1=162
ngram 2=747
ngram 3=1444

\1-grams:
-3.7839 <UNK>	0.0000
-0.8280 </s>	-2.6692
-0.8285 <s>	-1.1650
-3.7839 i	-0.2250
-1.9630 a	-1.0814
-3.7839 about	-0.2242
-3.2988 above	-0.1737
-3.7839 adjoining	-0.1932
-3.7839 after	-0.1788
-3.2988 all	-0.6117
-3.4749 almost	-0.2162
-3.1738 alone	-0.2880
-3.2988 also	-0.1777
-3.7839 am	-0.2250
-2.1738 and	-0.5033
-2.8728 another	-0.7396
-2.7345 are	-0.9268
-2.5998 at	-1.1579
-3.4749 away	-0.4007
-1.5998 back	-1.0947
-2.3780 behind	-0.8834
-3.7839 being	-0.2169
-3.4749 below	-0.1781
-2.8217 between	-0.2170
-3.0769 bit	-0.2779
-3.7839 blocking	-0.1788
-3.7839 board	-0.1552
-2.5998 bottom	-0.3207
-3.7839 bunch	-0.2174
-3.4749 but	-0.2107
-2.8728 by	-0.5307
-2.4537 centre	-0.3162
-3.4749 centered	-0.1876
-3.2988 close	-0.3187
-3.7839 closer	-0.2169
-2.5998 closest	-0.7439
-3.7839 cluster	-0.2174
-3.0769 color	-0.3519
-3.7839 colors	-0.2228
-3.2988 column	-0.6503
-1.3461 cone	-1.0214
-2.0200 cones	-0.7466
-2.3609 corner	-0.8793
-3.7839 deleted	-0.1552
-3.4749 diagonal	-0.1850
-3.0769 directly	-0.2681
-3.7839 down	-0.1552
-3.4749 east	-0.1146
-2.8728 edge	-1.0545
-3.7839 either	-0.2212
-3.7839 entire	-0.2250
-3.7839 extreme	-0.1907
-3.4749 face	-0.1548
-3.4749 facing	-0.2196
-2.8728 far	-0.4784
-3.4749 farthest	-0.1457
-3.7839 first	-0.2232
-3.7839 four	-0.1907
-3.7839 frame	-0.2228
-3.0769 from	-0.3849
-1.9434 front	-0.6385
-3.0769 further	-0.1669
-2.4335 furthest	-0.8257
-1.1198 green	-1.6012
-3.4749 greens	-0.1449
-3.0769 group	-0.3123
-3.7839 half	-0.2247
-2.2845 hand	-1.6759
-3.7839 hard	-0.1907
-3.2988 hardest	-0.1467
-3.1738 has	-0.2513
-3.4749 highest	-0.3950
-1.7004 in	-99.9990
-1.9835 is	-0.5430
-3.4749 isolated	-0.1562
-2.1738 it	-0.8542
-2.2988 it's	-0.5733
-3.4749 just	-0.2247
-3.2988 kinda	-0.3247
-2.9308 last	-0.3544
-1.7425 left	-0.5647
-3.7839 line	-0.1907
-3.0769 little	-0.3196
-3.7839 lone	-0.1907
-3.4749 lonely	-0.1146
-3.2988 lot	-0.6503
-2.6298 lower	-0.3920
-2.3445 lowest	-0.4657
-2.2574 middle	-0.2326
-2.8728 midline	-0.6836
-3.2988 most	-0.2158
-2.7759 near	-0.8693
-2.5718 next	-0.7179
-3.7839 no	-0.2250
-3.7839 non	-0.2250
-3.7839 north	-0.2249
-3.7839 not	-0.1788
-1.7589 of	-1.0816
-3.7839 off	-0.2235
-1.8516 on	-1.2188
-1.7385 one	-0.4928
-3.1738 ones	-0.1678
-2.8728 only	-0.2801
-3.2988 opposite	-0.3236
-3.4749 other	-0.1562
-3.2988 pair	-0.6503
-3.7839 part	-0.2174
-3.7839 previous	-0.2170
-3.7839 probably	-0.2178
-1.1506 purple	-1.6658
-3.4749 purples	-0.3312
-3.7839 rather	-0.2249
-3.2988 rear	-0.6503
-3.7839 rectangle	-0.2239
-1.7803 right	-0.4896
-2.5998 row	-0.6423
-2.5718 screen	-0.7337
-3.7839 second	-0.2247
-3.4749 see	-0.1458
-3.7839 shade	-0.2228
-3.0769 shades	-0.2073
-3.7839 shading	-0.1552
-3.4749 shadow	-0.1786
-3.1738 showing	-0.2242
-2.0599 side	-0.4692
-3.4749 sitting	-0.4008
-3.7839 slightly	-0.2169
-3.7839 somewhat	-0.2163
-3.7839 south	-0.2249
-3.7839 still	-0.2189
-2.8728 surrounded	-0.9537
-3.7839 surrounding	-0.2221
-3.7839 table	-0.1552
-3.7839 talking	-0.2250
-3.7839 than	-0.1788
-2.4335 that	-0.6779
-2.7759 that's	-0.3807
-0.9956 the	-0.8362
-3.7839 then	-0.2203
-2.7345 there	-1.1361
-2.6298 there's	-0.7167
-3.7839 third	-0.2247
-2.4749 this	-0.7790
-3.7839 those	-0.2247
-2.5455 three	-0.9116
-1.7345 to	-1.6863
-2.6620 top	-0.1657
-3.7839 touching	-0.2223
-2.8217 towards	-0.9537
-2.1961 two	-0.7746
-3.4749 up	-0.1547
-3.0769 upper	-0.5212
-3.2988 uppermost	-0.2941
-3.0769 us	-0.2952
-3.2988 very	-0.2228
-3.4749 was	-0.2174
-3.1738 way	-0.1954
-3.7839 we	-0.2249
-3.7839 west	-0.1907
-2.6620 which	-1.3344
-2.7345 with	-0.4459
-3.7839 you	-0.1552

\2-grams:
-3.3411 </s> </s> 2.4933
-0.0010 </s> <s> 0.0323
-3.3411 </s> green 0.1929
-0.9494 <s> </s> 0.5914
-3.3426 <s> a 0.0032
-2.5801 <s> another 0.2823
-0.9408 <s> back -1.0588
-3.3426 <s> blocking 0.0714
-2.5333 <s> centre -0.1714
-3.3426 <s> centered -0.0555
-3.3426 <s> cone -0.1033
-1.9494 <s> front -1.0582
-3.3426 <s> further -0.1170
-3.3426 <s> furthest -0.0223
-1.6941 <s> green 0.5160
-3.3426 <s> hard 0.0714
-2.8681 <s> hardest -0.0172
-3.3426 <s> has 0.0022
-2.3473 <s> in 0.2256
-3.3426 <s> isolated -0.0555
-2.8681 <s> it's -0.1425
-1.9951 <s> left -1.0060
-3.3426 <s> lonely -0.0555
-2.5801 <s> lower -0.1756
-3.3426 <s> middle -0.1379
-3.3426 <s> most -0.0907
-1.9494 <s> next -0.4538
-3.3426 <s> non 0.0714
-3.3426 <s> off 0.0714
-2.1712 <s> on 0.2430
-3.3426 <s> one -0.1443
-1.4053 <s> purple -0.8458
-3.3426 <s> right -0.1485
-3.3426 <s> south 0.0714
-0.6334 <s> the -0.2892
-2.8681 <s> there's -0.1006
-1.7453 <s> this -0.0129
-2.8681 <s> top -0.0588
-3.3426 <s> two 0.0460
-3.3426 <s> west 0.0714
-0.3932 I am 0.0714
-2.2128 a another 0.3235
-2.2128 a bunch 0.0714
-2.2128 a cluster 0.0714
-2.2128 a diagonal -0.0555
-0.4973 a green -1.0994
-1.7383 a group -0.0010
-1.0414 a little -0.4582
-1.4502 a lot 0.2247
-1.7383 a pair 0.3370
-0.5185 a purple -0.9570
-0.9165 a row 0.0828
-0.3932 about there -0.1032
-0.8703 above it -0.1413
-0.8703 above that -0.1291
-0.8703 above the -0.1533
-0.3932 adjoining purple -0.1261
-0.3932 after the -0.1519
-0.1078 all the -0.4316
-0.6942 almost in 0.3656
-0.9953 alone at 1.0505
-0.5208 alone in 0.1983
-0.9953 alone on 0.8676
-0.8703 also facing -0.0555
-0.8703 also the -0.1360
-0.8703 also with -0.0497
-0.3932 am talking 0.0714
-1.0212 and a 0.0768
-2.0165 and above -0.0907
-2.0165 and also -0.0907
-2.0165 and behind -0.0344
-2.0165 and closer 0.0714
-1.5420 and closest 0.4667
-1.5420 and green 0.5580
-0.7782 and in -0.7518
-2.0165 and it's -0.0634
-2.0165 and kinda -0.0907
-2.0165 and one -0.1520
-2.0165 and the -0.1486
-2.0165 and there's -0.1409
-0.4771 and to -0.1839
-2.0165 and towards 0.8464
-2.0165 and two 0.0460
-0.1761 another green -0.6497
-1.3475 another middle -0.1484
-0.8730 another purple 0.5707
-1.4346 are green -0.1269
-1.4346 are three 0.1474
-0.0872 are two -0.1504
-0.0280 at the -0.3405
-0.2198 away from -0.0770
-1.1405 back </s> 1.6055
-2.5751 back and 0.0225
-2.5751 back at 1.0505
-2.5751 back but -0.0555
-1.8125 back centre 0.0055
-1.3367 back corner -0.1127
-0.4337 back green -0.6522
-1.4037 back in 0.8427
-2.1006 back it -0.3086
-1.3367 back left -0.5572
-1.5798 back of -0.0521
-2.5751 back on 0.8676
-2.1006 back one -0.0586
-2.5751 back part 0.0714
-0.5191 back purple -0.4973
-2.5751 back right -0.0258
-2.5751 back top -0.1145
-1.0286 behind a -0.0180
-1.3167 behind and 0.1474
-1.7912 behind another 0.3235
-0.3565 behind it -0.1121
-1.3167 behind that -0.0060
-0.6198 behind the -0.0770
-0.3932 being to 0.8086
-0.6942 below another -0.0911
-0.6942 below the -0.0944
-1.3475 between a -0.1379
-1.3475 between it -0.0642
-1.3475 between purples 0.2475
-1.3475 between the -0.1486
-0.5382 between two -0.3520
-1.0922 bit further -0.1170
-0.6177 bit to 0.6412
-1.0922 bit towards 0.8464
-1.0922 bit up -0.0555
-0.3932 blocking the -0.1533
-0.3932 board </s> 2.4933
-0.4260 bottom </s> 1.8687
-1.1228 bottom a -0.2796
-1.5973 bottom and -0.1410
-1.5973 bottom it's -0.0287
-1.5973 bottom of 0.0641
-1.1228 bottom right -0.1932
-1.5973 bottom the -0.1360
-1.5973 bottom to 0.8086
-1.5973 bottom touching 0.0714
-0.3932 bunch of -0.1366
-0.6942 but on 0.8676
-0.6942 but to 0.8086
-1.2963 by a -0.1509
-1.2963 by four 0.0714
-0.8218 by green 0.2545
-0.3010 by purple 0.0371
-0.5643 centre </s> 1.8687
-1.2611 centre and -0.1366
-1.7356 centre behind -0.1321
-1.7356 centre cone -0.0851
-1.7356 centre front -0.1369
-1.7356 centre green 0.1929
-1.7356 centre line 0.0714
-1.2611 centre near 0.3780
-1.7356 centre of 0.0641
-0.7404 centre purple -0.0219
-1.7356 centre that's -0.0978
-1.7356 centre to 0.8086
-0.6942 centered and 0.0225
-0.6942 centered green -0.0251
-0.8703 close right -0.0641
-0.3958 close to 0.8235
-0.3932 closer to 0.8086
-1.0948 closest green -0.2840
-1.5693 closest purple -0.1195
-1.5693 closest the -0.1427
-0.1347 closest to 0.2443
-0.3932 cluster of -0.1437
-0.3296 color </s> 2.2137
-1.0922 color one -0.1443
-1.0922 color shades -0.1170
-0.3932 colors it's -0.0287
-0.1078 column of -0.1802
-0.5744 cone </s> 0.7836
-2.8310 cone I 0.0714
-2.3565 cone all 0.3370
-2.3565 cone alone 0.0701
-1.5347 cone and 0.1528
-1.4835 cone at 0.2499
-1.4835 cone behind 0.4717
-2.8310 cone being 0.0714
-2.8310 cone below -0.0555
-2.3565 cone between -0.1725
-2.8310 cone but -0.0555
-2.0684 cone directly -0.0890
-2.8310 cone further -0.1170
-1.5927 cone furthest -0.5626
-2.8310 cone greens -0.0555
-0.8357 cone in -99.9990
-0.9606 cone is -0.0941
-2.0684 cone it's -0.0536
-2.8310 cone kinda 0.0696
-2.8310 cone most -0.0907
-2.3565 cone near 0.6435
-2.8310 cone next -0.0813
-1.2073 cone on -0.2093
-2.3565 cone right -0.1255
-2.8310 cone second 0.0714
-2.3565 cone sitting 0.0801
-2.8310 cone somewhat 0.0714
-2.0217 cone surrounded 0.2553
-2.8310 cone that 0.0749
-2.8310 cone then 0.0714
-2.0684 cone there 0.3071
-2.8310 cone there's -0.1409
-0.9606 cone to -0.6225
-2.8310 cone towards 0.8464
-2.3565 cone which 0.7930
-1.5347 cone with 0.2677
-0.3158 cones </s> 1.1845
-2.1714 cones and -0.0744
-1.3621 cones behind -0.4472
-0.8751 cones in -99.9990
-2.1714 cones near 0.5454
-1.1761 cones on 0.6989
-2.1714 cones purple -0.0063
-2.1714 cones surrounding 0.0714
-2.1714 cones the -0.1533
-2.1714 cones there's -0.1145
-0.9331 cones to 0.1171
-2.1714 cones with -0.0497
-0.0889 corner </s> 1.3045
-1.8246 corner far -0.0033
-1.8246 corner in -0.0603
-1.8246 corner it's -0.1369
-1.3501 corner the -0.0835
-0.3932 deleted </s> 2.4933
-0.6942 diagonal of -0.1519
-0.6942 diagonal purple 0.1684
-0.6177 directly behind 0.0227
-1.0922 directly in -0.0603
-1.0922 directly opposite -0.0907
-1.0922 directly to 0.8086
-0.3932 down </s> 2.4933
-0.6942 east </s> 2.4933
-0.6942 east green 0.1929
-0.0512 edge </s> 1.7438
-1.3475 edge the -0.1360
-0.3932 either side 0.0266
-0.3932 entire frame 0.0714
-0.3932 extreme green 0.1929
-0.6942 face </s> 2.4933
-0.6942 face showing -0.1073
-0.6942 facing a -0.1122
-0.6942 facing towards 0.8464
-1.2963 far edge 0.8006
-1.2963 far in -0.0603
-0.5338 far left 0.1474
-0.5338 far right 0.0050
-0.6942 farthest </s> 2.4933
-0.6942 farthest to 0.8086
-0.3932 first behind -0.0344
-0.3932 four green -0.1269
-0.3932 frame it's -0.0287
-0.3296 from the -0.0200
-1.0922 from us 0.1206
-1.0922 from you 0.0714
-0.8388 front </s> 1.6469
-1.2368 front and -0.1624
-2.2321 front corner 0.5788
-2.2321 front far -0.1311
-0.7975 front green -0.2859
-1.4695 front in 0.7039
-2.2321 front left -0.0467
-0.4964 front of -0.2720
-2.2321 front on 0.8676
-1.2368 front one -0.1323
-2.2321 front pair 0.5043
-1.4228 front purple 0.3395
-2.2321 front right -0.1424
-2.2321 front row 0.2724
-2.2321 front to 0.8086
-1.0922 further back -0.1420
-1.0922 further behind 0.0982
-1.0922 further green -0.1269
-1.0922 further in -0.0603
-1.0922 further than 0.0714
-1.2804 furthest away 0.0801
-0.5836 furthest back 0.3920
-1.7549 furthest green -0.1469
-1.7549 furthest most -0.0907
-1.2804 furthest purple -0.3148
-0.3203 furthest to -0.0792
-0.2599 green </s> 0.1952
-3.0493 green and -0.1494
-2.5748 green between 0.0233
-0.5916 green cone 0.0830
-1.2247 green cones 0.2252
-3.0493 green furthest 0.1289
-2.0540 green it's -0.0163
-3.0493 green lowest -0.1473
-3.0493 green near 0.5454
-1.8110 green on 0.9363
-1.0879 green one -0.3104
-3.0493 green ones -0.1073
-3.0493 green that's -0.0978
-3.0493 green the -0.1486
-3.0493 green to 0.8086
-2.5748 green two -0.1001
-3.0493 green upper 0.1206
-0.6942 greens </s> 2.4933
-0.6942 greens in -0.1195
-0.2829 group of -0.3073
-0.3932 half way -0.1073
-0.6463 hand corner 0.6510
-1.8846 hand green 0.1929
-0.1296 hand side -0.0096
-0.3932 hard green -0.0251
-0.8703 hardest green 0.1929
-0.8703 hardest purple 0.1684
-0.8703 hardest to -0.1488
-0.9953 has a 0.0127
-0.9953 has the -0.1527
-0.5208 has three -0.1449
-0.2198 highest on 0.7002
-1.1213 in a -0.5037
-1.4735 in between -0.0220
-0.7138 in front -0.9389
-0.1565 in the -0.5244
-0.6301 is a -0.9010
-2.1994 is another -0.0911
-2.1994 is at 1.0505
-2.1994 is centered -0.0555
-2.1994 is close 0.0696
-1.4369 is closest 0.1408
-2.1994 is furthest -0.1302
-0.9611 is green -0.5101
-2.1994 is highest 0.2475
-0.8519 is in -0.4350
-2.1994 is lowest -0.1157
-1.7249 is on 0.7002
-2.1994 is one -0.1375
-1.4369 is purple -0.1087
-2.1994 is rather 0.0714
-2.1994 is slightly 0.0714
-1.7249 is surrounded 0.6333
-1.4369 is the -0.1439
-1.0280 is to 0.1841
-2.1994 is towards 0.8464
-2.1994 is very -0.0907
-0.6942 isolated green 0.1929
-0.6942 isolated purple -0.0063
-0.7304 it </s> 1.7438
-2.0267 it a 0.0127
-0.7304 it and -0.1278
-1.5522 it are 0.4194
-2.0267 it behind 0.0982
-1.5522 it has 0.0701
-0.7884 it is -0.7291
-1.2641 it there 0.8837
-0.7304 it there's -0.4559
-1.5522 it this 0.0499
-2.0267 it two -0.1053
-1.8984 it's a 0.0127
-1.8984 it's almost -0.0555
-1.4239 it's also -0.0172
-1.8984 it's below -0.0555
-0.7270 it's in -0.2589
-1.4239 it's near 0.6435
-1.8984 it's not 0.0714
-0.7270 it's on 0.2430
-1.8984 it's one -0.1375
-1.8984 it's purple 0.1684
-0.6021 it's the -0.1556
-1.8984 it's to 0.8086
-0.6942 just above -0.0907
-0.6942 just deleted 0.0714
-0.3958 kinda in 0.1983
-0.8703 kinda towards 0.8464
-1.2383 last cone -0.0191
-0.4758 last green -0.3962
-0.7638 last one 0.0304
-1.2383 last two -0.1490
-0.6612 left </s> 1.2667
-2.4346 left a -0.1122
-1.9601 left and 0.0894
-1.9601 left back -0.3005
-2.4346 left bottom 0.0505
-1.2632 left corner -0.1627
-1.9601 left edge 0.6333
-1.9601 left front -0.1250
-2.4346 left further -0.1170
-2.4346 left green 0.1929
-0.9274 left hand 0.8597
-1.9601 left in 0.1983
-2.4346 left left -0.1293
-0.8953 left of -0.9121
-1.9601 left on 0.7002
-1.0872 left purple -0.6322