A B C D E F G H I J K L M N O P Q R S T U V W X Y Z All
Doaa Mohey, El-Din
- Hierarchy Database Lexicon Solution for Sentiments Challenge
Authors
1 Information Systems Department, CU Cairo, EG
Source
Data Mining and Knowledge Engineering, Vol 9, No 2 (2017), Pagination: 29-33Abstract
Analyzing online sentiments challenges become hot research area to improve the accuracy and ease to understand. There are theoretical and technical sentiment challenges. World knowledge interprets old knowledge and updatable information. Till now, a few research in this challenge, because of the hardness of it and no standard measurement for it. This challenge clearly appears in two issues: recent events or linguistics similarities. This paper presents a new lexicon which is a solution for handling world knowledge challenge. This lexicon relies on a hierarchy database model. This research presents a new relationship between a topic domain and world knowledge challenge. Sentiment analysis plays a vital role in business decisions. Our target involves this importance in a scientific domain to support the researchers. The experiment concentrates on linguistics similarities and knowledge information not updatable. Its results achieve nearly 70%.
Keywords
Challenges, Explicit Negative, Implicit Negative, Reviews, Sentiment analysis, World Knowledge.- Negative Polarity Levels for Sentiment Analysis
Authors
1 Department of Information Systems, CU Cairo, EG
Source
Data Mining and Knowledge Engineering, Vol 9, No 1 (2017), Pagination: 24-28Abstract
The goal of Sentiment analysis is defining the attitude of a writer with respect to some topics or the overall sentiment polarity of a text, such as positive or negative. Online sentiments have a big effect in making a decision in business. There are several challenges in analyzing and evaluating sentiments. More than 60% of sentiments face a negative polarity challenge. In most research, negative consists of two levels: implicit and explicit. But we present new criteria for analyzing negative sentiments. The criteria include five negative levels that can effect on the word or sentence polarity. This paper proposes a new technique to improve the accuracy by analyzing negative reviews. We investigate the effect of evaluating negation in sentiment analysis based on word level. We also discuss the negative words and phrases types, with respect the conflict with several expressions. Our experimental results indicate that by evaluating and classifying for negative, precision relative to human ratings increases with 10%.