Gutenberg Children Books 2018-12-25-beta

LGE-clean corpus, min_word_count = 11; 500 clusters, varying agglomerative clustering parameters

Link Grammar 5.5.1, test_grammar updated 2018-10-19.
This notebook is shared as static Gutenberg-Children-Books-2018-12-25.html.
Output data shared via Gutenberg-Children-Books-2018-12-25 directory.

Basic settings

import os, sys, time
module_path = os.path.abspath(os.path.join('..'))
if module_path not in sys.path: sys.path.append(module_path)
from src.grammar_learner.utl import UTC, test_stats
from src.grammar_learner.read_files import check_dir, check_corpus
from src.grammar_learner.write_files import list2file
from src.grammar_learner.widgets import html_table
from src.grammar_learner.pqa_table import table_rows, params, wide_rows
tmpath = module_path + '/tmp/'
check_dir(tmpath, True, 'none')
start = time.time()
runs = (1,1)
print(UTC(), ':: module_path:', module_path)

2018-12-25 17:07:16 UTC :: module_path: /home/obaskov/94/language-learning

Corpus test settings

corpus = 'GCB' # 'Gutenberg-Children-Books-Caps' 
# corpus = 'CDS'
dataset = 'LG-E-clean'
kwargs = {
    'left_wall'     :   ''          ,
    'period'        :   False       ,
    'context'       :   1           ,
    'min_word_count':   11           ,
    'word_space'    :   'sparse'    ,
    'clustering'    :   ['agglomerative', 'ward', 'euclidean'],
    'clustering_metric' : ['silhouette', 'cosine'],
    'cluster_range' :   500         ,
    'top_level'     :   0.01        ,
    'grammar_rules' :   2           ,
    'max_disjuncts' :   1000000     ,   # off
    'tmpath'        :   tmpath      , 
    'verbose'       :   'log+'       ,
    'template_path' :   'poc-turtle',
    'linkage_limit' :   1000        }
rp = module_path + '/data/' + corpus + '/LG-E-clean/GCB-LG-English-clean.ull'
# rp = module_path + '/data/' + corpus + '/LG-E-clean/CDS-LG-English-clean.ull'
cp = rp  # corpus path = reference_path
runs = (1,1)
out_dir = module_path + '/output/' + 'Gutenberg-Children-var-ALE-par-' + str(UTC())[:10]
if check_corpus(rp, 'min'): print(UTC(), '\n', out_dir)

2018-12-25 17:07:16 UTC 
 /home/obaskov/94/language-learning/output/Gutenberg-Children-var-ALE-par-2018-12-25

Tests: min_word_count = 2; 500 clusters

%%capture
table = []
kwargs['clustering'] = ['agglomerative', 'ward', 'euclidean']
line = [['1.1', corpus, dataset, 0, 0, 'none']]
a, _, header, log, rules = wide_rows(line, out_dir, cp, rp, runs, **kwargs)
header[0] = 'Cell'
table.extend(a)

display(html_table([header] + a)); print(test_stats(log))

Cell

Corpus

Parsing

Space

Linkage

Affinity

G12n

Threshold

Rules

MWC

NN

SI

PA

PQ

F1

Top 5 cluster sizes

1.1

GCB

LG-E-clean

cALWEd

ward

euclidean

none

---

500

11

---

0.0

49%

44%

0.50

[516, 488, 472, 447, 394]

Cleaned dictionary: 8023 words, grammar learn time: 00:12:29, test time: 00:28:04 (hh:mm:ss)

%%capture
t2 = []
n = 2.0
for affinity in ['euclidean', 'manhattan', 'cosine']:
    n = round(n + 0.1, 1)
    kwargs['clustering'] = ['agglomerative', 'complete', affinity]
    line = [[n, corpus, dataset, 0, 0, 'none']]
    a, _, header, log, rules = wide_rows(line, out_dir, cp, rp, runs, **kwargs)
    t2.extend(a)
    table.extend(a)

display(html_table([header] + t2))

Line	Corpus	Parsing	Space	Linkage	Affinity	G12n	Threshold	Rules	MWC	NN	SI	PA	PQ	F1	Top 5 cluster sizes
2.1	GCB	LG-E-clean	cALCEd	complete	euclidean	none	---	500	11	---	0.0	56%	45%	0.48	[4657, 1684, 498, 211, 127]
2.2	GCB	LG-E-clean	cALCMd	complete	manhattan	none	---	500	11	---	0.0	56%	45%	0.48	[4657, 1684, 498, 211, 127]
2.3	GCB	LG-E-clean	cALCCd	complete	cosine	none	---	500	11	---	0.0	61%	41%	0.44	[206, 200, 143, 137, 111]

%%capture
t3 = []
m = 3.0
for affinity in ['euclidean', 'manhattan', 'cosine']:
    n = round(n + 0.1, 1)
    kwargs['clustering'] = ['agglomerative', 'average', affinity]
    line = [[n, corpus, dataset, 0, 0, 'none']]
    a, _, header, log, rules = wide_rows(line, out_dir, cp, rp, runs, **kwargs)
    t3.extend(a)
    table.extend(a)

display(html_table([header] + t3))

Grammar test for the last line (affinity = cosine) started 2018-12-26 ~10:31 UTC, connection lost, still running

Save results

display(html_table([header] + table))

print(UTC(), ':: finished, elapsed', str(round((time.time()-start)/3600.0, 1)), 'hours')
table_str = list2file(table, out_dir + '/all_tests_table.txt')
print('Results saved to', out_dir + '/all_tests_table.txt')