Affect analysis of text using fuzzy semantic typing

A hybrid approach to the sentiment analysis problem at the sentence level

“…In addition to the most common rules, a number of authors, among them [41,59,69], have pointed out the fact that having rules for negation handling and to deal with the use of specific PoS particles, like 'but, despite, unless, ...' could positively affect the final outcome of a classification exercise.…”

Main Concepts, State of the Art and Future Research Questions in Sentiment Analysis

“…It is interesting to see that in the cases of China and India, and to a certain extent Europe as well, research efforts are present in both camps (ML and FS). However, for the USA, at least for the period of time chosen and the search keywords and data sources utilised, the focus is clearly on the ML camp, despite the fact that one of the most influential papers supporting the use of FS was written in 2001 in the USA [57]. One may think as well that two of the most reputed researches in SA/OM, Bing Liu and Bo Pang, have made ML their fundamental tool.…”

“…In the literature, the use of Fuzzy Logic is found in Anaphora Resolution -given an expression S i , its interpretation depends upon another expression S j in context -in the work of Witte & Bergler [71]. Analysis of affect in text using fuzzy sets by means of the concept of fuzzy semantic typing can be found in [57]. Named Entity recognition has been addressed as well using fuzzy techniques in [32].…”

“…In 2001, Subasic & Huettner [57] released the most important contribution we are aware of to the use of Fuzzy Sets/Logic principles in SA. Since then, others have followed their footsteps, but certainly, Fuzzy Sets Theory has been so far a bit of an outsider in the research field for SA.…”

An Evaluation Framework and Adaptive Architecture for Automated Sentiment Detection

“…In another work, Wiebe and Riloff [23] produce training data for the training of a Naive Bayes subjectivity classifier by employing a rule-based classifier. Subasic and Huettner [19] apply fuzzy methods to analysing affect in writings. Blitzer et al [2] present an approach using similarities between differing domains in order to adapt a sentiment classifier to a new domain.…”