This is how the text analysis work
The text analysis is AI-based and utilizes OpenAI's latest model, GPT-4, in combination with triggers based on Brilliant's method and experience.
Triggers, or prompts, are used to guide the AI model in the right direction and ensure relevant insights.
A few fundamentals:
- The analysis is based on all text comments provided
- Areas frequently mentioned are categorised into different topics
- The respondent's feeling towards the topic discussed is categorised by sentiment
The example below demonstrates how the analysis picks up different parts of a text comment.
The text analysis identifies two topics here: simplicity and information.
In addition to the two topics being discussed, there are also sentiments associated with these topics. Here, the sentiment analysis finds a positive sentiment towards simplicity and a negative sentiment associated with information.
The sentiment associated with specific topics is reported in the form of a sentiment score.