Aspect Based Sentiment Analysis | The Linguists Studio

ASPECT-BASED

SENTIMENT ANALYSIS

Aspect based Sentiment Analysis is also known as Feature or Attribute based sentence Analysis. ABSA is used to analyse different features/attributes/aspects of a target feature. Sentiment analysis is an important task in natural language understanding and has a wide range of real-world applications. The typical sentiment analysis focus on predicting the positive or negative polarity of the given sentence(s). This task works in the setting that the given text has only one aspect and polarity. A more general and complicated task would be to predict the aspects mentioned in a sentence and the sentiments associated with each one of them. This generalized task is called aspect-based sentiment analysis (ABSA).

document level

Document-level Sentiment Analysis, analyses text and the context in which it is situated. Depending on what the text is about, ABSA at document level informs whether a whole document, message, etc, is overall positive or negative. Depending on what variables are assessed, it is possible to have an output of a numerical score of the sentiment value.

sentence level

Sentence-level Sentiment Analysis, is a deeper and thorough form of sentiment analysis in a body of text. When applying sentence level ABSA, we can detect the sentiment of each sentence within the document to sieve out specific sentiments and sentiments which would be classified in a broad category such as 'positive' or 'negative'. Other than a numerical score, it is possible to use the results in a interpretable and actionable manner in whatever context it is situated in.

topic based

Topic based Sentiment Analysis will identify not only positive or negative phrases, but also the specific topic that is being referred as positive or negative. This will also identify when several opinions are made in the same sentences, like in “While I love Apple, the latest iPhone is not so great”: in this case two opinions will be detected, one with a positive value (about Apple) and one with a negative (or not-so-positive) value (about the latest iPhone).