Re: [Corpora-List] Chi-Square

From: Joaquim Francisco Ferreira da Silva (jfs@di.fct.unl.pt)
Date: Sun Sep 17 2006 - 13:14:26 MET DST

  • Next message: Marco Baroni: "Re: [Corpora-List] Chi-Square"

    Crayton Walker wrote:

    > A simple question about statistical measures.
    >
    > Could someone explain in very simple terms why we don’t normally use
    > Chi-square as a measure of collocational significance? We tend to use
    > t-score and MI and not Chi-square. Why not? I am not a mathematician
    > so would appreciate it if you could keep it simple.
    >
    > Many thanks
    >
    > Crayton Walker
    >
    > University of Birmingham
    >
    I have used Chi-Square. Results are good. The reason why it is not so
    used is, I think, it is because it is not so easy to implement. However,
    Chi-square results are very similar to other two very
    simple-to-implement measures:

    SCP(a,b)= f(a,b)^2 / (f(a) * f(b))

    or

    Cosine(a,b)= f(a,b) / (sqrt(f(a)) * sqrt(f(b))

    Joaquim



    This archive was generated by hypermail 2b29 : Sun Sep 17 2006 - 13:34:16 MET DST