SentiWORDNET is a lexical resource designed to aid in opinion mining, a field that focuses on extracting and analyzing user opinions from text. The resource associates each WordNet synset with three numerical scores (Obj(s), Pos(s), Neg(s)) that describe the objective, positive, and negative connotations of the terms within the synset. The development of SentiWORDNET involves quantitative analysis of synset glosses and semi-supervised classification using a committee of eight ternary classifiers. These classifiers are trained using seed sets of positive, negative, and objective synsets, and their results are combined to produce the final scores. SentiWORDNET is freely available and includes a web-based graphical user interface for visualizing the scores. The resource aims to provide a more nuanced understanding of term connotations, which can be particularly useful in applications such as sentiment analysis and customer relationship management.SentiWORDNET is a lexical resource designed to aid in opinion mining, a field that focuses on extracting and analyzing user opinions from text. The resource associates each WordNet synset with three numerical scores (Obj(s), Pos(s), Neg(s)) that describe the objective, positive, and negative connotations of the terms within the synset. The development of SentiWORDNET involves quantitative analysis of synset glosses and semi-supervised classification using a committee of eight ternary classifiers. These classifiers are trained using seed sets of positive, negative, and objective synsets, and their results are combined to produce the final scores. SentiWORDNET is freely available and includes a web-based graphical user interface for visualizing the scores. The resource aims to provide a more nuanced understanding of term connotations, which can be particularly useful in applications such as sentiment analysis and customer relationship management.