Social media posts classification system
View/ Open
Date
2019Author
Atwine, Nickson
Mugabo, Amuza
Mulidwa, Henry
Turyahiirwa, Happiness
Metadata
Show full item recordAbstract
This report demonstrates the production of a sentiment analysis system, with the following main objectives set:
1. Implement a machine learning algorithm to perform sentiment analysis.
2. Understand and implement natural language processing technique.
3. Achieve a classification accuracy of over 75%.
4. Build a graphical user interface to enhance the interaction with the users of the system as well as manage users of the system and also for visualization purpose.
In order to produce the software artefacts presented in the report, computer science knowledge, as well as machine learning and natural language processing techniques were employed. Consequently, the concepts and techniques, which contributed to the development of the project, such as the Naive Bayes algorithm, logistic regression, are explained.
Furthermore, a high-level view and a low-level view of the system produced are detailed in subsequent chapters. The quality is assessed in the Achievements section, where a performance benchmark and other evaluation techniques are employed.