correlation measure of 2 tables

Discussion in 'Technical Analysis' started by zedDoubleNaught, May 21, 2011.

  1. Is there a correlation measure for two tables of data?
    My 2 tables are stop x limit, with data from 2 separate procedures. I would like to find out if they are related. One option I could do is find the regular linear correlation row by row. But if there was a better statistic measure, I'd be interested to learn about and use that.
  2. Another idea I had -- I have QT Octave, perhaps I could make 3-D mesh plots of the 2 tables, then see how much the match / differ. Is there a 3-D regression, like linear regression but in 3 dimensions?
  3. <a href="'s_chi-square_test"> Chi square </a>
  4. In Octave, corr(Table1( : ),Table2( : )) will return the
    Pearson Correlation btween the two tables.

    A better measure might be Distance Correlation. There is an R package (energy) for it, not sure if there is anything for Octave/Matlab.
  5. Thanks -- that should give me a number sufficient for what I'm looking for, and especially great it's just just one command.