Proposed aspect extraction algorithm for Arabic text reviews

Abstract EN

Opinion mining from reviews is a very crucial area in NLP.

This area has many applications in social networks, business intelligence, and decision making.

Aspect extraction is the main step to achieve opinion mining.

This paper proposed an algorithm for aspect extraction from reviews in the Arabic language, to determine the aspects that the reviewers are described in their comments.

The proposed algorithm begins with analyzing the comments dataset using latent Dirichlet analysis (LDA) to identify the aspects and its essential representative words, then extracting nouns and its' adjectives as a possible aspect phrase in a review.

After that the categorizing process to categorize the extracted phrases according to the words specified from LDA analysis.

The proposed approach has been tested by using two standard Arabic reviews datasets.

The result was auspicious in spite of the difficulties if the Arabic natural language processing