Normal görünüm MARC görünümü ISBD görünümü

Utilizing text classification methods to monitor criminal activities on Twitter Omar Al-Debagy; Supervisor: Devrim Seral

Yazar:

Al-Debagy, Omar

Katkıda bulunan(lar):

Supervisor: Seral, Devrim

Dil: İngilizce Yayın ayrıntıları:Nicosia Cyprus International University 2015Tanım: XIV, 81 p. table, figure 30.5 cm CDİçerik türü:

text

Ortam türü:

unmediated

Taşıyıcı türü:

volume

Konu(lar):

Özet: 'ABSTRACT The purpose of this research is to find the most efficient text classification techniques to monitor and track criminal activities on Twitter that are written in Arabic. In this research, we compared the performance of different classifiers with different preprocessing options. Classification made using three different classes, which they are supporting ISIS (Islamic State of Iraq and Syria), opposing ISIS, and neutral tweets which they are unrelated to any of the two other classes. The classifiers that we used in this study are Naïve Bayes, Support Vector Machine, k-Nearest Neighbors, and Random Forests, with different preprocessing options like TF/IDF, stemming, stop words removal, and normalization. So we concluded that the most efficient classifier for this research will be Naïve Bayes with accuracy of 88% by utilizing 3000 tweets as training set and 1200 tweets as test set. '

Materyal türü:

Thesis

Mevcut ( 1 )
Başlık notları ( 3 )

Mevcut
Materyal türü	Geçerli Kütüphane	Koleksiyon	Yer Numarası	Durum	Notlar	İade tarihi	Barkod	Materyal Ayırtmaları
Thesis	CIU LIBRARY Tez Koleksiyonu	Tez Koleksiyonu	YL 609 A43 2015 (Rafa gözat(Aşağıda açılır))	Kullanılabilir	Information Systems Engineering Department		T668

Toplam ayırtılanlar: 0

CIU LIBRARY raflarına göz atılıyor, Raftaki konumu: Tez Koleksiyonu, Koleksiyon: Tez Koleksiyonu Raf tarayıcısını kapatın(Raf tarayıcısını kapatır)

Önceki	Kullanılabilir kapak resmi yok	Kullanılabilir kapak resmi yok	Kullanılabilir kapak resmi yok	Kullanılabilir kapak resmi yok	Kullanılabilir kapak resmi yok	Kullanılabilir kapak resmi yok	Kullanılabilir kapak resmi yok	Sonraki
Önceki	YL 606 I27 2015 Mechanical properties of waste clay brick based geopolymer cured at various temperature	YL 607 U34 2015 Microeconomic and macroeconomic effect on bank sector profitability Evidence from Nigeria	YL 608 H97 2015 The burnout levels of Kurdish EFL teachers working in governmental schools in Duhok	YL 609 A43 2015 Utilizing text classification methods to monitor criminal activities on Twitter	YL 61 A35 2005 İznik gölü'nde yaşayan levkit balığı (Rutilus frisii nordmann, 1840) populasyonu üzerine bir araştırma	YL 610 I27 2015 Error analysis of the written english essays by the second year students at the university of Raparin in Northern Iraq	YL 611 A43 2015 Technical evaluation of class f fly ash based geopolymer	Sonraki

Includes CD

Includes references (76-81 p.)

'ABSTRACT The purpose of this research is to find the most efficient text classification techniques to monitor and track criminal activities on Twitter that are written in Arabic. In this research, we compared the performance of different classifiers with different preprocessing options. Classification made using three different classes, which they are supporting ISIS (Islamic State of Iraq and Syria), opposing ISIS, and neutral tweets which they are unrelated to any of the two other classes. The classifiers that we used in this study are Naïve Bayes, Support Vector Machine, k-Nearest Neighbors, and Random Forests, with different preprocessing options like TF/IDF, stemming, stop words removal, and normalization. So we concluded that the most efficient classifier for this research will be Naïve Bayes with accuracy of 88% by utilizing 3000 tweets as training set and 1200 tweets as test set. '

Ayırt
Yazdır
Cihaza gönder
Künyeyi kaydet
BIBTEX Dublin Core ISBD MARCXML MODS (XML) RIS
Başka yerde arama

Bu eser adını şu dizinde arayın:
Other Libraries (WorldCat) Other Databases (Google Scholar) Online Stores (Bookfinder.com) Open Library (openlibrary.org)