Karteikarten
Week 10 - Sentiment Analysis
Karten in diesem Stapel (4)
- 1What is Sentiment AnalysisComputationally identifying and categorizing emotions expressed in a text especially as positive, negative or neutral
- 2How does sentiment analysis works in terms of some features/functions?1) Identify if text is opinionated or contains subjective communiction
2) convert the piece(s) of texxt to a feature vector so it can be measured
3) Part-of-speech (POS) tagging for word sense disambiguation
4) detect negation (eg 'I don't hate this book')
5) Syntax analysis & modality identification (description, opinion etc?) - 3What are 2 challenges associated with sentiment analysis?The fact that a piece of text needs to be classified into one of two polarized categories or locate its position on the continuum between these two polarities
Identifying and correctly assessing semantic embedding - 4How are deceptive opinions handled in Spam detection?A corpus is created for both deceptive and true opinions. Both are analysed and applied to training of the Spam system using:
Text categorization: use n-gram based classifiers
Psychological deception detection: look for negative emotion, distancing
Genre Identification: imaginative vs informative writing
Informationen
- Beschreibung
- 1 Artificial Intelligence in the Marketplace (Week 10 - Sentiment Analysis) Karteikarten am Week 10 - Sentiment Analysis, erstellt von Daniel Gomez-Windshuttle am 10/12/2015.
- Erstellt von
- Daniel Gomez-Windshuttle
- Aufrufe
- 0
- Bewertung