Implicit Aspect Extraction in Product Reviews Using FIN Algorithm

Abstract

Online transactions are growing very rapidly right now. Every online transaction is often accompanied by a review. Product reviews from buyers can be used by sellers as feedback. Product reviews provide information as a consideration for decision making for potential buyers to find out the strengths and weaknesses of the product. Identifying specific product features from reviews written by buyers becomes a solution to make it easier to find information. Aspect-based extraction in sentiment analysis is divided into two, explicit aspects and implicit aspects. The explicit aspect is the explicit aspect in the sentence while the implicit aspect is the aspect that is implied in the sentence. The extraction carried out in this study is based on implicit aspects to determine its features because the majority of existing studies extract explicit aspects. Implicit extraction aspects of product reviews using the FIN algorithm in association rule mining. The dataset is in English text where to extract features using TF-IDF and select features using Particle Swarm optimization. Selected features are grouped using k-Means. After features are grouped based on their value, an associative rule is made using the FIN algorithm. The minimum support value applied and the number of sentence variations cause the accuracy value obtained by 0.678.