I came across this recent KDD paper titled as Concepts-Bridges: Uncovering Conceptual Bridges Based On Biomedical Concept Evolution.

In page 6 of the paper they have used Chi-square to obtain a ranked list.

This make me questioned how to use chi-square for ranking as I initially thought that it is a hypothesis testing approach. Is it like ranking using p-values?

I also has a ranking problem where I would like to use chi-square to rank my terms.

With the curiosity of knowing how chi-square is used in ranking, my question is how to use chi-square to get a ranked list using a simple example.

My preferred language is python. However, I am happy to receive answers using r as well.


I still could not find a solution for this. The current answer for this question looks promising. However, the answer does not provide details how to implement this. It would be really great if you could help me to resolve my problem. My preferred language is python. However, I am fine in getting answers in other languages too.


I still could not find a way to do this. Please help me.

  • 1
    $\begingroup$ Link does not work. $\endgroup$ – user2974951 Sep 6 '19 at 7:31
  • $\begingroup$ @user2974951 thank you for pointing that out.I edited the question :) $\endgroup$ – Elly Sep 6 '19 at 9:47
  • $\begingroup$ mlwiki.org/index.php/Chi-Squared_Ranking - google helped me =) by value of $\chi^2$ $\endgroup$ – German Demidov Oct 5 '19 at 9:18
  • $\begingroup$ @GermanDemidov thanks a lot for the helpful comment. Do you know how to implement this in python? I would like to write a pythonprogramme that does the similar. Thank you :) $\endgroup$ – Elly Oct 5 '19 at 11:24

In the paper "Concepts-Bridges: Uncovering Conceptual Bridges Based On Biomedical Concept Evolution" where they mentioned about the chi-squared test, they reference the paper "Mining Biomedical Knowledge Using Mutual information ABC". In this paper, we can see the table to understand how they used chi-squared distribution.

They talk about the bridge between "fish oils" and "Raynaud disease". As Raynaud disease causes restricted blood flow, the blood viscosity play a role and we can see in the table that this has relatively high chi-square value. As higher chi-square value suggests a low p-value and hence that the two are not independent(null hypothesis assumes they are independent). So, the ranking is based on the chi-square value going from high for dependent variables to low for independent variables.

| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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