Sentiment mining using immune network algorithm /Raja Muhammad Hafiz Raja Kamarudin

With the abundant of knowledge accessible via Internet and the likes, many have ventured into using that knowledge in order to understand deeply on how the available knowledge affected the whole world directly or indirectly. Sentiment analysis or opinion mining is an application that tries to identify and extract subjective information from source materials. Many business applied sentiment analysis in their marketing and others practical usage. In short, sentiment analysis has become an important aspect in our today’s life with or without we realizing it. Currently, there are many techniques available in order to harvest necessary sentiment values. However, there are very few researches have been done with the intention to utilize biological-inspired computing in sentiment analysis. This research tries to utilize biological-inspired computing algorithm, Immune Network algorithm, in this sentiment analysis field. This research also studies the effect of preprocessing on the data which has shown an unexpected result, the data with no-preprocessing done produces more accurate result than the data with preprocessing albeit very small in percentage, < 2%. The results obtained by utilizing Immune Network in sentiment mining are not very impressive compared to other Machine Learning algorithms. The algorithms can be improved furthermore by including other disciplines such as Natural Language Processing (NLP) or others. However, this is a step stone towards developing a biological-inspired Sentiment Mining algorithm