Arabic Opinion Mining Using Lexicon Based Classification And Machine Learning Algorithms (in Arabic)
|
Author |
|
|
Published in |
April 2014 |
|
Abstract |
Sentiment Analysis and Opinion Mining became an attractive field for a large number of researchers in the field of Natural Language Processing and Web Mining. This was encouraged by the large amount of opinions and impressions available on social networking sites and e-commerce sites, which are easily accessible materials compared to the period preceding the Internet. This research aims to test the efficiency of two document level opinion mining methodologies on the Arabic language: the Lexicon Based method, and Machine Learning algorithms. In the first Methodology, we build an effective primary system used to determine polarity of Arabic articles, some assumptions were used, and many resources were built, varying from adjectives and verbs lists extracted from OCA corpus and other was extracted from the Internet and Arabic dictionary. An opinion mining system was built in order to test the efficiency. It performs different linguistic analysis steps in order to extract texts’ features for opinion classification. In the second Methodology, many linguistic processes were performed. The efficiency of three classification algorithms SVM, KNN, NB was tested. Keywords: Opinion mining, Sentiment Analysis, Information Retrieval, Machine Learning, Natural Language Processing, Classification. |
|
Link to read full paper |
