Abstract In this work, we propose a framework suitable for multidimensional data mining based on tensor. A Tensor Data Model (TDM) provides us with high order data structure and naive description for information retrieval. Among others, we discuss multidimensional rule mining here. Generally, association rule mining (or extraction of association rules) concerns about co-related transaction records of single predicate, and hard to examine the... In this work, we propose a framework suitable for multidimensional data mining based on tensor. A Tensor Data Model (TDM) provides us with high order data structure and naive description for information retrieval. Among others, we discuss multidimensional rule mining here. Generally, association rule mining (or extraction of association rules) concerns about co-related transaction records of single predicate, and hard to examine the ones over multiple predicates since it takes heavy time- and space- complexities. Here we show TDM allows us to model several operations specific to multidimensional data mining yet to reduce amount of description. Read More Read Less

Abstract With the thriving of review sites on the Web, people have begun to express their opinions on a wide variety of products on several services. Sentiment analysis on entities of products thus becomes a rapid and effective way of assessing public opinion for business marketing. Sentiment analysis is the computational study of people’s opinions, attitude, emotion or appraisal pertaining to topics,... With the thriving of review sites on the Web, people have begun to express their opinions on a wide variety of products on several services. Sentiment analysis on entities of products thus becomes a rapid and effective way of assessing public opinion for business marketing. Sentiment analysis is the computational study of people’s opinions, attitude, emotion or appraisal pertaining to topics, objects, products, services, organizations, individuals, events or any attributes of them. Aspect detection in sentiment analysis helps customers to effectively navigate product information by zooming in on the product features they are interested in. In this paper we introduce Sentiment Miner, a novel unsupervised framework for aspect detection problem in a sentiment analysis system for online customer reviews. Sentiment Miner tries to detect aspects using a bootstrapping algorithm based on pointwise mutual information measure. The proposed framework considers multiword aspects as atomic terms by utilizing C-value method. Experimental results show that the Sentiment Miner improves the precision, recall and the F-score, and outperforms the state-of-the-art baseline approach. Read More Read Less