mclm
The goal of mclm is to gather various functions in support of quantitative corpus linguistics. It contains classes for corpus files, frequency lists, association scores dataframes and concordances and functions to create them, manipulate them and read them from and write them to files.
The package is a companion to the Methods in Corpus Linguistics course at the Advanced Master in Linguistics (KU Leuven), but can be used for basic corpus linguistic analyses. In particular, it offers a number of learnr tutorials on how to perform basic tasks with mclm and filter objects with PERL-flavor regular expressions.
Installation
You can install the development version of mclm from GitHub with:
remotes::install_github("masterclm/mclm")Examples
Below are some basic usages of mclm.
The freqlist() function can generate a frequency list from either the
text of a corpus or corpus files.
library(mclm)
#> Loading required package: ca
#> Loading required package: tibble
#>
#> Attaching package: 'mclm'
#> The following object is masked from 'package:tibble':
#>
#> as_data_frame
toy_corpus <- "Once upon a time there was a tiny toy corpus.
It consisted of three sentences. And it lived happily ever after."
flist <- freqlist(toy_corpus, as_text = TRUE)
print(flist, n = 5)
#> Frequency list (types in list: 19, tokens in list: 21)
#> rank type abs_freq nrm_freq
#> ---- --------- -------- --------
#> 1 a 2 952.381
#> 2 it 2 952.381
#> 3 after 1 476.190
#> 4 and 1 476.190
#> 5 consisted 1 476.190
#> ...The get_fnames() function creates a list of filenames based on the
contents of a directory and can be given to different functions that
process corpora. surf_cooc(), for example, computes the surface
co-occurrences
of an item, such as the type “government”, in a given corpus. These
co-occurrences can be provided to assoc_scores() to compute the
association strength of different collocates of the node (here
“government”) in the corpus.
corpus_files <- get_fnames(system.file("extdata", "cleveland", package = "mclm"))
length(corpus_files)
#> [1] 4
surf <- surf_cooc(corpus_files, "government", w_left = 5, w_right = 5)
assoc_scores(surf)
#> Association scores (types in list: 77)
#> type a PMI G_signed| b c d dir exp_a DP_rows
#> 1 the 230 0.578 39.554|1321 2152 20276 1 154.072 0.052
#> 2 of 136 0.403 11.259|1415 1454 20974 1 102.844 0.023
#> 3 to 57 0.286 2.323|1494 666 21762 1 46.765 0.007
#> 4 by 39 1.017 17.223|1512 259 22169 1 19.275 0.014
#> 5 in 37 0.038 0.028|1514 520 21908 1 36.028 0.001
#> 6 this 37 1.811 45.360|1514 126 22302 1 10.543 0.018
#> 7 and 36 -0.634 -8.873|1515 828 21600 -1 55.885 -0.014
#> 8 a 28 0.207 0.600|1523 347 22081 1 24.256 0.003
#> 9 has 18 1.238 11.232|1533 100 22328 1 7.632 0.007
#> 10 be 15 -0.332 -0.927|1536 277 22151 -1 18.887 -0.003
#> 11 that 15 -0.067 -0.036|1536 228 22200 -1 15.718 0.000
#> 12 for 14 -0.185 -0.258|1537 232 22196 -1 15.912 -0.001
#> 13 with 14 0.136 0.130|1537 183 22245 1 12.742 0.001
#> 14 their 13 0.112 0.082|1538 173 22255 1 12.031 0.001
#> 15 which 10 -0.120 -0.076|1541 158 22270 -1 10.867 -0.001
#> 16 as 9 -0.128 -0.078|1542 143 22285 -1 9.832 -0.001
#> 17 made 9 1.393 6.903|1542 44 22384 1 3.428 0.004
#> 18 our 9 -0.297 -0.440|1542 162 22266 -1 11.061 -0.001
#> 19 states 9 0.491 1.012|1542 90 22338 1 6.403 0.002
#> 20 been 8 0.169 0.114|1543 102 22326 1 7.115 0.001
#> ...
#> <number of extra columns to the right: 7>The function conc() finds occurrences of a regular expression in a
corpus and generates a concordance.
conc(corpus_files, "govern")
#> Concordance-based data frame (number of observations: 29)
#> idx left|match |right
#> 1 ...heir power and right of self-|govern|ment they have committed to o...
#> 2 ... the strength and safety of a|govern|ment by the people. In each s...
#> 3 ...d the surest guaranty of good|govern|ment. But the best results in...
#> 4 ...results in the operation of a|govern|ment wherein every citizen ha...
#> 5 ...efits which our happy form of|govern|ment can bestow. On this ausp...
#> 6 ...ation of a republican form of|govern|ment and most compatible with...
#> 7 ...f. In the administration of a|govern|ment pledged to do equal and ...
#> 8 ... benefits of the best form of|govern|ment ever vouchsafed to man. ...
#> 9 ...hina. The admitted right of a|govern|ment to prevent the influx of...
#> 10 ...asure of that sovereign self-|govern|ment pertaining to the States...
#> 11 ...his land of freedom, of self-|govern|ment, and of laws, here peace...
#> 12 ... of successful constitutional|govern|ment, maintenance of good fai...
#> 13 ...ulty pending with any foreign|govern|ment. The Argentine Governmen...
#> 14 ...itation in favor of a foreign|govern|ment upon the right of select...
#> 15 ... several States into a single|govern|ment. In these contests betwe...
#> 16 ... and complications of distant|govern|ments. Therefore I am unable ...
#> 17 ...hina. The admitted right of a|govern|ment to prevent the influx of...
#> 18 ...Kongo has been organized as a|govern|ment under the sovereignty of...
#> 19 ...he plenipotentiaries of other|govern|ments, thus making the United...
#> 20 ...purpose toward their original|govern|ments. These evils have had m...
#> 21 ...the safety and welfare of any|govern|ment. Emergency calling for a...
#> 22 ...es at legations. Some foreign|govern|ments do not recognize the un...
#> 23 ...he President shall invite the|govern|ments of the countries compos...
#> 24 ... attitude and intent of those|govern|ments in respect of the estab...
#> 25 ...ioned that the views of these|govern|ments are in each instance su...
#> 26 ...to the fixed rules which must|govern|the Army, I am inclined to ag...
#> 27 ...ected by a republican form of|govern|ment, to which they owe alleg...
#> 28 ...nd the people who desire good|govern|ment, having secured this sta...
#> 29 ...g for the use of the District|govern|ment which shall better secur...
#>
#> This data frame has 6 columns:
#> column
#> 1 glob_id
#> 2 id
#> 3 source
#> 4 left
#> 5 match
#> 6 right