SEMANTIC Recurrence:.


97 views
Uploaded on:
Category: News / Events
Description
SEMANTIC Recurrence: another take a gander at word recurrence checks What are the present issues with word recurrence numbers? (especially in the field of corpus phonetics) Definition or build of "word" – Gardner (2007) Structure versus importance (Read 2000)
Transcripts
Slide 1

SEMANTIC FREQUENCY: another take a gander at word recurrence numbers

Slide 2

What are the present issues with word recurrence checks? (especially in the field of corpus phonetics) Definition or develop of “word” – Gardner (2007) Form versus significance (Read 2000) Polysemy and Homonomy: high recurrence words are the most polysemous (Ravin and Leacock 2000) Idiosyncratic nature of word structures and implications Computer-produced rundown regularly strip expressions of their implications and just number structures

Slide 3

Implications: particularly identified with the field of English as a Second Language Word records – what English is it speaking to? Scope – what is the rate\'s truth of representation in the dialect? Assertiveness and learnability – what are the mental substances of showing and learning vocabulary and significance?

Slide 4

RESEARCH QUESTIONS What improvement does structure based recurrence tallies versus semantically-construct recurrence include have light of the accompanying: 1. PC produced word records utilized for pedagogical purposes? 2. assessments of word scope of agent English writings, both composed and talked? 3. evaluations of lexical requests on dialect learners (the learning weight of a word)?

Slide 5

Methodology of the Current Study Purpose: perform an agent test semantic recurrence number of lemmas to answer the accompanying examination questions

Slide 6

Methods 1. Haphazardly chose 46 lemmas from the BNC with 1500 or more aggregate events (method of reasoning for utilizing lemmas quickly) 2. Discovered a broad (however not comprehensive) rundown of faculties for every lemma from WordNet Modified the sense list by conflating some polysemous faculties and including a few detects that were experienced while rating semantic recurrence in connections from the BNC (the objective was to attempt and take out polysemy and keep up homonomy)

Slide 7

Lemmas versus Word Families Lemmas – just incorporate a word and its intonations (for the most part a grammatical feature) Ex. work/lives up to expectations/working/worked Word families – incorporate a word, its affectations and straightforward (or firmly related) inductions Ex. From RANGE program (Nation) Active/Actively/Activities/Activity/Inactive/Inactivity/Activist/Activists/Activism

Slide 8

46 Randomly Selected Lemmas

Slide 9

Example faculties from WordNet FAIR – descriptive word 1. S: (adj) reasonable , simply (free from partiality or self-premium or predisposition or double dealing; adjusting with set up principles or tenets) "a reasonable referee"; "fair deal"; "on a reasonable footing"; "a reasonable fight"; "by reasonable means or foul" 2. S: (adj) reasonable , fairish , sensible (not unnecessary or compelling) "a fairish income"; "reasonable prices" 3. S: (adj) bonny , bonnie , attractive , reasonable , sightly (exceptionally satisfying to the eye) "my bonny lass"; "there\'s a bonny straight beyond"; "a attractive face"; "young reasonable maidens" 4. S: (adj) reasonable ((of a baseball) hit between the foul lines) "he hit a reasonable ball over the third base bag" 5. S: (adj) normal , reasonable , unremarkable , ordinary (lacking outstanding quality or capacity) "a novel of normal merit"; "only a reasonable execution of the sonata"; "in reasonable health"; "the bore of the understudies has gone from average to above average"; "the execution was mediocre at best" 6. S: (adj) reasonable (alluringly female) "the reasonable sex" 7. S: (adj) clean , reasonable ((of an original copy) having couple of changes or remedies) "fair copy"; "a clean manuscript" 8. S: (adj) legit , reasonable (picked up or earned without duping or taking) "an legitimate wage"; "an reasonable penny" 9. S: (adj) reasonable (free of mists or downpour) "today will be reasonable and warm" 10. S: (adj) reasonable , fairish ((utilized of hair or skin) pale or light-shaded) "a reasonable complexion";

Slide 10

Semantic Relatedness Scale Nagy and Anderson (1984) (utilized for the conflation process)

Slide 11

Example of conflated faculties FAIR – descriptor S: (adj) reasonable , simply (free from preference or self-premium or inclination or trickery; acclimating with built up measures or guidelines) "a reasonable referee"; "fair deal"; "on a reasonable footing"; "a reasonable fight"; "by reasonable means or foul" + S: (adj) genuine , reasonable (picked up or earned without bamboozling or taking) "an legitimate wage"; "an reasonable penny" - identified with definition #4 in the feeling of adjustment to administer; ‘fair enough’ = simply, okay, adequate, great, fine; reasonable play (inside of the standards) S: (adj) reasonable , fairish , sensible (not intemperate or amazing) "a fairish income"; "reasonable prices” ; can mean more towards a considerable measure or toward an extensive sum yet not to the complete great or exorbitance S: (adj) bonny , bonnie , attractive , reasonable , sightly (extremely satisfying to the eye) "my bonny lass"; "there\'s a bonny inlet beyond"; "a attractive face"; "young reasonable maidens" + S: (adj) reasonable (appealingly ladylike) "the reasonable sex" - identified with definition 10 (CS8) contingent upon social standards S: (adj) reasonable ((of a baseball) hit between the foul lines) "he hit a reasonable ball over the third base bag" –somewhat identified with 1 in similarity of standard S: (adj) normal , reasonable , average , mediocre (lacking extraordinary quality or capacity) "a novel of normal merit"; "only a reasonable execution of the sonata"; "in reasonable health"; "the bore of the understudies has gone from average to above average"; "the execution was mediocre at best" – to some degree identified with definition 2 in being in the center as opposed to at the extremes S: (adj) clean , reasonable ((of a composition) having couple of modifications or amendments) "fair copy"; "a clean manuscript" S: (adj) reasonable (free of mists or downpour) "today will be reasonable and warm" 8. S: (adj) reasonable , fairish ((utilized of hair or skin) pale or light-hued) "a reasonable complexion";

Slide 12

Literal versus Non-literal “the figurative utilization of lion (e.g., John is a lion) is liable to be dealt with as ‘the same word,’ while the solid and allegorical employments of crane (‘kind of bird’ and ‘machine for lifting substantial objects’) are more inclined to be dealt with as free words and in this way individuals from distinctive lemmas. On the off chance that it is hard to gathering words implications under headwords at the conceptual level of the lexicon, it is a great deal more hard to allocate words in writings unambiguously to their lemmas.” (Knowles and Mohd Don 2004: 70)

Slide 13

Methods (cont’d) 4. Examined the connections and evaluated (or appointed a sense to) 100 talked and 100 composed settings (200 aggregate) for every lemma. **Each lemma was twofold evaluated. Triple appraisals were done when errors between evaluations happened. Faculties for every lemma were tallied up and rates figured in Excel.

Slide 14

Example Context Ratings

Slide 15

RESULTS QUESTION #1 What improvement does structure based recurrence numbers versus semantically-construct recurrence checks have with respect to the accompanying: 1. PC created word records utilized for pedagogical purposes? *How much homonomy exists in the lemmas?

Slide 16

Break Down of Sense Distributions LEMMA + POS s1 s2 s3 s4 s5 s6 s7 s8 s9 s10 s11 s12 s13 s14 s15 s16

Slide 17

Sense Distributions (cont’d) LEMMA + POS s1 s2 s3 s4 s5 s6 s7 s8 s9 s10 s11 s12 s13 s14 s15 s16

Slide 18

Effect on Word Lists? Lemma Form-based Frequency Semantically-based Frequency

Slide 19

Conclusions The presence of homonomy in these 23 words makes 58 aggregate detects - more than a multiplying the quantity of potential high recurrence lemmas The lemmas of high recurrence word records would be changed and extended in their request of lemmas, substance, and size so as to incorporate every single critical lemma Apparent particular nature of vocabulary - 23 of the words had almost no homonomy demonstrating that some high recurrence lemmas have implications that are exceptionally unsurprising, steady, and in this manner prone to be adapted all the more effectively

Slide 20

RESULTS Question #2 What improvement does structure based recurrence checks versus semantically-construct recurrence numbers have with respect to the accompanying: 2. appraisals of word scope of delegate English writings, both composed and talked? *How a great part of the dialect are every lemma ACTUALLY covering?

Slide 21

Word Coverage and Comprehension Thresholds 95% cognizance of words with a specific end goal to grasp a content or talk 98% understanding for joy perusing (Nation 2006)

Slide 22

5 most homonymous lemmas

Slide 23

Coverage

Slide 24

Extrapolations

Slide 25

Written versus Talked

Slide 26

Writing versus Talking

Slide 27

Conclusions The presence of homonomy makes more lemmas which: a) diminishes the measure of scope that every lemma speaks to in the dialect b) expands the quantity of lemmas that must be found out with a specific end goal to achieve the 95% understanding limit Differences between recurrence of lemmas in talked and composed dialects makes a requirement for independent high-recurrence word records for these two noteworthy registers

Slide 28

RESULTS Question #3 What improvement does structure based recurrence tallies versus seman

Recommended
View more...