What do you think about this photo? A novel approach to opinion and sentiment analysis of photo comments

We propose a practical unsupervised approach to opinion and sentiment analysis of photo comments with a real-valued strength orientation. We extract two types of opinions: opinions that relate to the photo quality and general sentiments targeted towards objects depicted on the photo. Our approach combines linguistic features for part of speech tagging, traditional statistical methods for modelling word importance in the photo comment corpus (in a real-valued scale), and a predefined lexicon for detecting negative and positive opinion orientation. In addition, we apply a semi-automatic photo feature detection method and introduce a set of syntactic patterns to resolve opinion references. The results of our user study among 49 non-expert participants of different ages showed no statistical differences between user evaluation and the algorithm.