Users can mark articles that are returned by Kbot as “useful” or “not useful”. This feedback builds a knowledge base called the Recommender system. It offers useful articles to users when appropriate, speeding up the search. This page describes the knowledge training, the process of managing the Recommender system.

Concept

When users ask Kbot to find a solution, it uses search systems, such as Google, Stack Overflow, etc. Although some results returned by Kbot prove to be a perfect match, others might be less helpful.

Users provide their feedback on articles by marking them “Useful” or “Not useful”. It builds a knowledge base called Recommender system. As this knowledge base is built upon articles evaluated by users, it is unique for each Kbot instance.

To manage the Recommender system an expert performs the following:

• Reviewing the articles that users have marked either useful or not useful.

• Training the bot.

Reports

Sentence evaluation sandbox

Enter a user statement to see the articles that Kbot will return.

Text similarity evaluation

Enter two user statements for Kbot to calculate the similarity between them.