Cross Validated is a question and answer site for people interested in statistics, machine learning, data analysis, data mining, and data visualization. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I've a little question regarding the Discounted Cumulated Gain (DCG) (Sorry, I couldn't find the papers of Järvelin and Kekäläinen). Can this evaluation-metric be used when a information retrieval system has only a binary classification of relevance? If this topic is misplaced, please tell where I should post it.

Thank you

share|improve this question
up vote 0 down vote accepted

Sure it can. Using the standard logarithmic discount it then allows you to conclude - for instance - that a search result with relevances 0,1,1 and 0 is slightly better than a search result with relevances 1,0,0 and 1.

As a side note, note that the logarithmic discount is not motivated by theory; it is merely used out of convenience and with other discounts your ranking of the results will differ. It may be worth while to think and pick a scheme that captures what it actually is you want to optimise.

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.