Title :
Author recognition by Abstract Feature Extraction
Author :
Yasdi, Murat ; Diri, Banu
Author_Institution :
Bilgisayar Muhendisligi Bolumu, Yildiz Tek. Univ., Istanbul, Turkey
Abstract :
The purpose of this study is to show the success of Abstract Feature Extraction Method in multi dimensional feature vectors studies. Author recognition study is taken as an application area and word root and 2 gram´s are chosen as feature vectors. The success of the Abstract Feature Extraction method in classification is shown on both Turkish and English data sets by comparing with feature extraction methods such as PCA, CFS, chi-square.
Keywords :
feature extraction; natural language processing; pattern classification; CFS; English data sets; PCA; Turkish data sets; abstract feature extraction method; author recognition; chi-square; multidimensional feature vectors; word root; Abstracts; Bayesian methods; Computers; Feature extraction; Niobium; Principal component analysis; Text categorization;
Conference_Titel :
Signal Processing and Communications Applications Conference (SIU), 2012 20th
Conference_Location :
Mugla
Print_ISBN :
978-1-4673-0055-1
Electronic_ISBN :
978-1-4673-0054-4
DOI :
10.1109/SIU.2012.6204690
