Sentiment defines if the mention has positive, negative or neutral overtone.
In SentiOne, we use both statistical based (machine learning algorithms) and lexicon based classification solutions. Our sentiment analysis is based on research work by John R. Crawford and Julie D. Henry. They analyzed the Positive and Negative Affect Schedule (PANAS). Based on their research, we have created algorithms that help us determine the emotional attitude of the author to discussed subject.
PANAS - Positive and Negative Affect Schedule
The Positive and Negative Affect is used to study the frequency and intensity of the experience of positive and negative affective states, which constitute one of the basic components of happiness and subjective well-being. Negative people see the world in darker colors. NA and PA scales reflect aspects of the disposition of a subject. High NA points to the subjective anguish and unpleasant obligation, and low NA to lack of experience. In contrast, PA shows when a person experiences a pleasant engagement with the environment. Thus, emotions such as enthusiasm and alertness are indicators of high PA, while lethargy and sadness are characterized by low PA. So in one sentence - when someone uses expressions that indicate negative emotions, it is highly probable that this statement is negative and vice versa.
The PANAS consists of 10 statements that relate to positive emotional states and 10 negative ones. This 20-unit questionnaire was invented by Messrs Watson, Clark and Tellegen in 1988.
Updated about 2 years ago