Roberto Tonino

Information Retrieval: Classic model

Posted on:

Note: the content of this post is mainly notes taken from professor Stefano Mizzaro’s lessons on (Web) Information Retrieval at Uniud.

Classic information retrieval model

What this means

There is an initial set of documents which are then indexed, forming a set of indexed documents.

There is an user having an information need. This information need is expressed by a query inserted into the IR system.

Once the query is inserted, the system matches it with the indexed documents and produces a set of retrieved documents. If the set of documents doesn’t satisfy the information need, the user reformulates the query and repeats the process.

Main problem

The information need is not static! It changes with every loop of the model because the user benefits of the results to improve it or completely change it, in the situation the initial one was not actually the real information need. This makes the classic model quite trivial and mainly wrong.

Martia Bates created a term that fits this situation, berry picking: berries do not come all together – they are scattered in bushes and need to be picked up one at a time