Mining for subjectivity from text is a hard science requiring machine learning based approaches to harness information from the varied and large data. Subjectivity is all about finding opinions, affects, and sentiments from texts. This could take the form of processing blogs, reviews, tweets, editorials, and general articles as well as several other sources of textual content. Subjectivity is important in a wide range of domains. It can be valuable for calculating and identifying particular trends and forecasting for real-world applications such as in financial markets, fashion, events, economic indicators, social markers, political voting, chart toppers, and a lot more where understanding attitudes and feelings matter on a particular 'thing' or 'concept'. Automatically finding such information is hard and utilizing machine learning is an area of research analysis for identifying and extracting opinions and sentiments from texts. In order to achieve and develop such an application requires understanding and applying statistical natural language processing.
Sources of information on sentiment analysis are hard to find as well as the subject is relatively new. The below links are a valuable entry point towards further information.
Subjectivity Analysis
sentiment-analysis
opinion mining sentiment analysis survey
Twitter-as-a-Corpus-for-Sentiment-Analysis-and-Opinion-Mining
sentiwordnet
twitter-sentiment-analysis-training-corpus-dataset-2012-09-22
sentiment140
