Notes‎ > ‎

Notes on Bob

posted 10 Feb 2012, 09:18 by David Sherlock
Bobs code will be pretty basic, I think one think I’ve learnt is that staying basic is a key to many ABMs. Bob inputs stuff from the web and things about whats happening. 

Condidtional probability is the probability of an event happening as a result of having observed a different event. A basic way to work out what the conditional probability of something is the number of times Event A and Event B happen together, divided by how many times Event B happens on its own. (A & B) / (B). This is all Bob will do. The code looks something like what IBM have going on here.


<?php
/**
* Returns conditional probability of $A given $B and $Data.
* $Data is an indexed array.  Each element of the $Data array 
* consists of an A measurement and B measurment on a sample 
* item.
*/
function getConditionalProbabilty($A, $B, $Data) {
  $NumAB   = 0;
  $NumB    = 0;
  $NumData = count($Data);
  for ($i=0; $i < $NumData; $i++) {
    if (in_array($B, $Data[$i])) {
      $NumB++;
      if (in_array($A, $Data[$i])) {
        $NumAB++;
      }
    }
  }
  return $NumAB / $NumB;
}
?>

Bob will operate on counts and frequencies. This isn’t good because we want to probability values rather than frequency counts.  I haven’t decided if I’ll expand on Bobs intelligence or create a new agent. I need to get the data in first! I’m tied and its Friday. 
Comments