This thesis was written as a part of the MSc in ICT Systems at the International Hellenic University.
With the rising importance of knowledge interchange, along with the emergence of Web 2.0 an important problem has been introduced. This problem is namely how social content can be annotated with relevant semantic information, if this problem is solved in a distributed manner it will allow for the analysis of user made content to be more easily made in an automatic fashion. The analysis of social interactions within communities begins with search and without any further knowledge about the domain on which is being searched the analysis is limited.
In the thesis a methodology is defined for efficient and accurate sentiment analysis across the social networking platform Twitter. To add sentiment to the searching process ontologies are automatically (or semi-automatically) created in an object-attribute pattern. The methodology consists of three steps. (i) Create a domain ontology; (ii) Sentiment analysis on Tweets; (iii) Quantitative analysis on outputted sentiment scores.
The primary intention was to improve how sentiment analysis on Twitter will be conducted in a practical sense. This is achieved by allowing the inclusion of more data in searching and more relevant search into the Twitter API, due the use of ontologies.
Furthermore the thesis identifies a number of promising areas for future work. Finally, it gives a comprehensive overview of related, similar and subsumed approaches.
I would like to express my sincere thanks to Dr. E. Kontopoulos, my supervisor, for hi encouragement and support during my research. He always remained approachable and offered the right amount of guidance at all times. Also I would like to thank Dr C. Berberidis who directed me to pursue such an interesting dissertation topic.
Collections
Show Collections