الفهرس 
Chapter 1: IntroductionOnly 14 pages are availabe for public view
 |  | from | 90 |  |  |
| | | from | 90 | | |
Abstract
In the last few years, online reviews where individuals express their thoughts, interests, experiences and opinions have majorly spread over the internet. Sentiment analysis field of study has evolved to analyze these online reviews and provide valuable insights for both individuals and organizations that may help them in making decisions. Unfortunately, the performance of sentiment analysis process is affected by the nature of online reviews’ content that may contain emoticons and negation words. Moreover, spam reviews have been written for the purpose of deceiving others. These spam reviews may greatly influence online marketing and prevent both individuals and organizations from concluding real ideas about certain services or products. Therefore, there is a need to develop an approach that considers these issues.
In this thesis, an enhanced approach for sentiment analysis is proposed which aims to enhance the performance of classifying reviews based on their features and assigning an accurate sentiment score to each feature. This enhanced approach is achieved by handling negation, detecting emoticons, and detecting spam reviews using a combination of different types of properties which leads to achieving better predictive performance. Moreover, this approach examines the impact of using three different feature extraction methods on the performance of sentiment classification which are extracting all nouns, extracting only the nouns that occur frequently, and extracting frequent nouns by applying Apriori algorithm.
Several experiments have been carried out to validate the effectiveness of the proposed approach. The performance of the proposed approach has been measured using different types of evaluation metrics which are accuracy, precision, recall, and f1 score. The proposed approach has been verified against three datasets of different sizes. The experimental results showed the efficiency of the proposed approach in detecting spam reviews, classifying reviews based on their features and assigning an accurate sentiment score to each feature. The proposed approach achieves a maximum accuracy of about 99.06% in detecting spam reviews and outperforms the existing related works with an average value of 13.35% for accuracy. The proposed approach achieves as well a maximum accuracy of about 97.13% in classifying reviews after considering the three main challenges: negation handling, emoticons detection, and spam reviews detection together and after employing “extracting frequent nouns by applying Apriori algorithm” as a feature extraction method, where there is an improvement in accuracy value of about 29.72%, and a great saving in the feature space by 96.9% versus when not considering these three main challenges together along with this feature extraction method.