東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

Dialect Identification Using Natural...

Djuve, Kari Oline.

FindBook

Google Book

Amazon

博客來

Dialect Identification Using Natural Language Processing and Machine Learning.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Dialect Identification Using Natural Language Processing and Machine Learning./
作者:	Djuve, Kari Oline.
出版者:	Ann Arbor : ProQuest Dissertations & Theses, : 2018,
面頁冊數:	95 p.
附註:	Source: Masters Abstracts International, Volume: 80-05.
Contained By:	Masters Abstracts International80-05.
標題:	Web Studies. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10979751
ISBN:	9780438667976

Dialect Identification Using Natural Language Processing and Machine Learning.
Djuve, Kari Oline.

Dialect Identification Using Natural Language Processing and Machine Learning. - Ann Arbor : ProQuest Dissertations & Theses, 2018 - 95 p.

Source: Masters Abstracts International, Volume: 80-05.

Thesis (M.S.)--Southeastern Louisiana University, 2018.

This item must not be sold to any third party vendors.

Dialectology is the linguist field that focuses on the study of regional variances among speakers of a language. Although linguists primarily use spoken data for dialect studies, written language also shows marked differences between members of different dialectal groups. This thesis study aims to use a supervised decision tree classifier to predict the dialect class of some written texts. Since the context in which the writer is writing in-i.e. formal letter versus journal entry-plays a major role in the style used by writers, a data set of texts pulled from the social media platform Twitter will be used. Tweets, the messages users post on Twitter, are short and informal, thus very likely to be written in the users' dialect. Ten thousand tweets are pulled from two dialectal regions-Southern American English and New England American English dialects. However, these texts are also noisy. This study goes over how to normalize, or clean, these noisy texts so that Natural language processing (natural language processing) tools can be used for feature selection. Common feature selection methods will be applied to the texts to select the most information rich features while reducing the number of features that are redundant or too common to be useful in differentiating between the different dialect classes. Two feature selection pipelines are applied for comparison. The Feature Selection 1 pipeline performs punctuation, non-alphabetic tokens, and stop-words removal. The Feature Selection 2 pipeline extends the Feature Selection 1 pipeline by adding a stemmer. Additionally, TF-IDF weighting will be used for comparison. After the feature selection pipelines, a decision tree classifier is used to make class predictions.

ISBN: 9780438667976Subjects--Topical Terms:

1026830
Web Studies.

Dialect Identification Using Natural Language Processing and Machine Learning.
LDR:02782nmm a2200325 4500 001 2206192
005 20190828133647.5
008 201008s2018 ||||||||||||||||| ||eng d
020 $a 9780438667976
035 $a (MiAaPQ)AAI10979751
035 $a (MiAaPQ)selu:10515
035 $a AAI10979751
040 $a MiAaPQ $c MiAaPQ
100 1 $a Djuve, Kari Oline. $3 3433086
245 1 0 $a Dialect Identification Using Natural Language Processing and Machine Learning.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2018
300 $a 95 p.
500 $a Source: Masters Abstracts International, Volume: 80-05.
500 $a Publisher info.: Dissertation/Thesis.
500 $a Burris, John.
502 $a Thesis (M.S.)--Southeastern Louisiana University, 2018.
506 $a This item must not be sold to any third party vendors.
520 $a Dialectology is the linguist field that focuses on the study of regional variances among speakers of a language. Although linguists primarily use spoken data for dialect studies, written language also shows marked differences between members of different dialectal groups. This thesis study aims to use a supervised decision tree classifier to predict the dialect class of some written texts. Since the context in which the writer is writing in-i.e. formal letter versus journal entry-plays a major role in the style used by writers, a data set of texts pulled from the social media platform Twitter will be used. Tweets, the messages users post on Twitter, are short and informal, thus very likely to be written in the users' dialect. Ten thousand tweets are pulled from two dialectal regions-Southern American English and New England American English dialects. However, these texts are also noisy. This study goes over how to normalize, or clean, these noisy texts so that Natural language processing (natural language processing) tools can be used for feature selection. Common feature selection methods will be applied to the texts to select the most information rich features while reducing the number of features that are redundant or too common to be useful in differentiating between the different dialect classes. Two feature selection pipelines are applied for comparison. The Feature Selection 1 pipeline performs punctuation, non-alphabetic tokens, and stop-words removal. The Feature Selection 2 pipeline extends the Feature Selection 1 pipeline by adding a stemmer. Additionally, TF-IDF weighting will be used for comparison. After the feature selection pipelines, a decision tree classifier is used to make class predictions.
590 $a School code: 1559.
650 4 $a Web Studies. $3 1026830
650 4 $a Computer science. $3 523869
690 $a 0646
690 $a 0984
710 2 $a Southeastern Louisiana University. $b Interdisciplinary Science and Technology: College of Science and Technology. $3 3286808
773 0 $t Masters Abstracts International $g 80-05.
790 $a 1559
791 $a M.S.
792 $a 2018
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10979751