Lexical approach to sentiment analysis on hotel reviews | |
| Author | Bajracharya, Yemini |
| Call Number | AIT Caps. Proj. no.CS-15-04 |
| Subject(s) | Natural language processing (Computer science) Lexicology Data processing |
| Note | A Capstone Project report submitted in partial ful llment of the requirements for the degree of Bachelor of Science in Computer Science, School of Engineering and Technology |
| Publisher | Asian Institute of Technology |
| Series Statement | Caps. Proj. ; no. CS-15-04 |
| Abstract | Sentiment Analysis is the study of peoples' opinions and sentiments in written form. It has been one of the important topics in research industry due to two main reasons. Firstly, people are in uenced by online opinions and reviews before consuming any product or service. Online opinions are found in abundance and need to be tracked constantly. Secondly, there has not been any research on this topic before the year 2000, and challenging problems still exist. The reason for this is that the available opinionated text in digital form was sparse before. The escalating growth in social media on the web has extended the text processing research work not only in com- puter science but also in business management and society as a whole. Currently, the main focus of sentiment analysis is on improving existing opinion mining resources. In this project, the main goal is to explore the limitations of an online lexical resource, SentiWordNet, in terms of lexical approaches. I focused on a sample selection of 2000 reviews from TripAdvisor and created hand- tagged lexicon. This project includes a description of rules for negation identi cation and calculation.I compared the lexicon along with the technique of automated negations was compared with SentiWordNet lexicon, which proved to be more accurate. I also compared the overall polarity of the reviews using both, the Hand-tagged lexicon and SentiWordNet with the negation applied. This brought signi cant improvement and was the most accurate of the methods I tested. SentiWordNet is a valuable lexicon and has been used by many researchers for sentiment analysis. This study presents the results of applying the SentiWordNet lexical resource using negation to improve its performance. I conclude the investigation with interesting results in calculating polarities. |
| Year | 2015 |
| Corresponding Series Added Entry | Asian Institute of Technology. Caps. Proj. ; no. CS-15-04 |
| Type | Project |
| School | School of Engineering and Technology (SET) |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Computer Science (CS) |
| Chairperson(s) | Choochart Haruechaiyasak; |
| Examination Committee(s) | Dailey, Matthew;Guha, Sumanta ; |
| Degree | Capstone Project (B. Sc.) - Asian Institute of Technology, 2015 |