The widespread of social communication media in the Web has produced a large volume of opinionated textual data stored in digital format. Social media constitutes a rich source for sentiment analysis and understanding of the opinions spontaneously expressed. Many scientific proposals have arisen in the last years aiming to deal with sentiment analysis issues. However, most of them do not address both spatial and temporal dimensions that enable a more accurate analysis and a better understanding

more »

... of the mood of people when using social media. In this context, we rely on Geographical Information Retrieval techniques in order to infer geographical locations mentioned in Twitter microtexts (tweets). We propose an approach based on two well-known classification algorithms for detecting the sentiment polarity on tweets considering both spatial and temporal information. Our approach differs from related work since it does not rely on Part-Of-Speech (POS)Taggers. The proposed approach is evaluated through a case study using a dataset of Portuguese Twitter microtexts harvested during a big event which took place in Brazil. The achieved results not only outperformed related work as they have shown which is possible to perform sentiment analysis with a good accuracy even without relying on POS-Taggers