Php logics, correlations

[fandreacci]

//SAMPLE 1 DATA
$x[] = 1301.45;
$x[] = 1028.28;
$x[] = 2386.98;
$x[] = 479.66;
$x[] = 4017.31;
$x[] = 2947.39;
$x[] = 1561.90;
$x[] = 2510.86;
$x[] = 1808.22;
$x[] = 987.93;
$x[] = 2658.48;
$x[] = 1110.40;
$x[] = 465.75;
$x[] = 1782.85;
$x[] = 1512.25;
$x[] = 1748.03;
$x[] = 1003.62;
$x[] = 578.13;
$x[] = 2103.99;
$x[] = 6249.16;
$x[] = 743.12;
$x[] = 789.02;
$x[] = 457.58;
$x[] = 540.98;
$x[] = 449.46;
$x[] = 500.44;
$x[] = 1181.20;
$x[] = 1422.96;
$x[] = 4493.26;
$x[] = 673.63;
$x[] = 664.99;
$x[] = 3598.93;
$x[] = 1652.60;
$x[] = 1887.02;
$x[] = 2109.16;
$x[] = 3190.90;
$x[] = 2025.76;
$x[] = 1786.44;
$x[] = 2065.75;
$x[] = 2160.60;
$x[] = 709.32;
$x[] = 632.55;
$x[] = 522.68;
$x[] = 782.89;
$x[] = 830.31;
$x[] = 699.99;
$x[] = 599.99;
$x[] = 417.27;
$x[] = 235.12;
$x[] = 243.44;
$x[] = 235.12;
$x[] = 578.14;
$x[] = 686.87;
$x[] = 460.73;
$x[] = 343.43;
$x[] = 347.80;
$x[] = 521.70;
$x[] = 1135.06;
$x[] = 808.50;
$x[] = 273.84;
$x[] = 848.52;
$x[] = 1298.37;
$x[] = 517.25;
$x[] = 345.53;
$x[] = 1108.37;
$x[] = 1101.49;
$x[] = 432.19;
$x[] = 2665.15;
$x[] = 4211.44;
$x[] = 432.19;
$x[] = 3674.36;
$x[] = 3115.58;
$x[] = 2387.14;
$x[] = 2843.54;
$x[] = 2337.72;
$x[] = 1187.00;
$x[] = 1121.48;
$x[] = 2193.11;
$x[] = 1759.18;
$x[] = 2392.31;
$x[] = 1657.36;
$x[] = 1652.35;
$x[] = 1279.14;
$x[] = 2141.22;
$x[] = 1824.40;
$x[] = 2588.53;
$x[] = 1147.30;
$x[] = 1047.59;
$x[] = 1453.04;
$x[] = 1116.75;
$x[] = 725.48;
$x[] = 371.68;
$x[] = 425.70;
$x[] = 796.70;

//SAMPLE 2 DATA
$y[] = 1538.44;
$y[] = 4162.40;
$y[] = 890.86;
$y[] = 935.39;
$y[] = 2123.34;
$y[] = 794.65;
$y[] = 1205.56;
$y[] = 2726.06;
$y[] = 1222.14;
$y[] = 1520.81;
$y[] = 1342.54;
$y[] = 3836.90;
$y[] = 698.63;
$y[] = 3647.81;
$y[] = 671.83;
$y[] = 1851.27;
$y[] = 3232.88;
$y[] = 5082.69;
$y[] = 3221.18;
$y[] = 1846.11;
$y[] = 3685.61;
$y[] = 9521.83;
$y[] = 3309.90;
$y[] = 5828.21;
$y[] = 8243.15;
$y[] = 3082.19;
$y[] = 2630.29;
$y[] = 698.63;
$y[] = 753.56;
$y[] = 602.91;
$y[] = 1000.40;
$y[] = 1041.48;
$y[] = 1753.65;
$y[] = 411.20;
$y[] = 3439.41;
$y[] = 4074.20;
$y[] = 3179.73;
$y[] = 6746.77;
$y[] = 2311.18;
$y[] = 3107.68;
$y[] = 2464.01;
$y[] = 2957.40;
$y[] = 879.64;
$y[] = 815.64;
$y[] = 1864.97;
$y[] = 976.47;
$y[] = 1010.01;
$y[] = 989.93;
$y[] = 976.47;
$y[] = 589.34;
$y[] = 729.22;
$y[] = 968.86;
$y[] = 912.29;
$y[] = 1205.43;
$y[] = 675.79;
$y[] = 452.81;
$y[] = 476.22;
$y[] = 392.89;
$y[] = 822.52;
$y[] = 835.76;
$y[] = 1026.29;
$y[] = 809.29;
$y[] = 1469.14;
$y[] = 739.07;
$y[] = 1839.42;
$y[] = 3748.92;
$y[] = 2664.70;
$y[] = 847.59;
$y[] = 3190.39;
$y[] = 4695.29;
$y[] = 4005.25;
$y[] = 1437.19;
$y[] = 4128.00;
$y[] = 4085.50;
$y[] = 1309.04;
$y[] = 4533.39;
$y[] = 3809.06;
$y[] = 749.13;
$y[] = 3474.56;
$y[] = 2657.05;
$y[] = 2629.93;
$y[] = 2714.28;
$y[] = 4233.91;
$y[] = 2215.41;
$y[] = 1145.49;
$y[] = 1474.78;
$y[] = 1286.16;
$y[] = 1337.04;
$y[] = 2222.28;
$y[] = 2621.11;
$y[] = 2571.50;
$y[] = 3223.01;
$y[] = 2759.02;
$y[] = 3045.74;
$y[] = 2493.63;
$y[] = 1134.24;
$y[] = 1308.62;
$y[] = 1212.77;
$y[] = 3093.44;
$y[] = 1428.42;

//SIMPLE CORRELATION FUNCION
function simplecorrelation($x, $y) {

  if(count($x) != count($y)) {
    print "series must have same size";
return false;
  }
  
  // Number of values or elements 
  $n = count($x);
  
  //Sum of the product of first and Second Scores
  $sumPxy = 0;
  for($i=0;$i<count($x);$i++) {
    $sumPxy += $x[$i]*$y[$i];  
  }
  
  //Sum of First Scores
  $sumx = 0;
  //Sum of square First Scores
  $sumxs = 0;
  
  //Sum of Second  Scores
  $sumy = 0;
  //Sum of square Second  Scores
  $sumys = 0;
  
  for($i=0;$i<count($x);$i++) {
    $sumx += $x[$i];
    $sumxs += $x[$i]*$x[$i];
    $sumy += $y[$i];
    $sumys += $y[$i]*$y[$i];	  
  }
  
  $correlation = ((($n*$sumPxy)-($sumx*$sumy)) / sqrt(($n*$sumxs - ($sumx*$sumx))*($n*$sumys - ($sumy*$sumy))));
  return $correlation;
}

//TEST
$a = array(60,61,62,63,65);
$b = array(3.1,3.6,3.8,4,4.1);

print simplecorrelation($a,$b);

//IT RETURNS 0.91187249533773

 

Ok, it works. However my big problem is that my $x and $y have different sizes.

 

$x has 94 as $y has 100 values.

 

I must write a code that removes the difference (in this case, remove 6 values from $y) giving the best correlation as possible. So the code must not remove 6 random elements, the code must remove the 6 which removed will give to me the best correlation (using my symplecorrelation()) between $x and $y.

 

Got it?

 

Thanks!

 

 

 

[fandreacci]

I did it!

 

however it's non functional to large data. Can someone help me to improve?

 

my total number of combations $total = $num * $num; is 10000, but i'm getting duplicate values that uses the array space. And if I increase the $total value, php can't handle with the large data.

 

//SIMPLE CORRELATION FUNCION
function simplecorrelation($x, $y) {

  if(count($x) != count($y)) {
    print "series must have same size";
return false;
  }
  
  // Number of values or elements 
  $n = count($x);
  
  //Sum of the product of first and Second Scores
  $sumPxy = 0;
  for($i=0;$i<count($x);$i++) {
    $sumPxy += $x[$i]*$y[$i];  
  }
  
  //Sum of First Scores
  $sumx = 0;
  //Sum of square First Scores
  $sumxs = 0;
  
  //Sum of Second  Scores
  $sumy = 0;
  //Sum of square Second  Scores
  $sumys = 0;
  
  for($i=0;$i<count($x);$i++) {
    $sumx += $x[$i];
    $sumxs += $x[$i]*$x[$i];
    $sumy += $y[$i];
    $sumys += $y[$i]*$y[$i];	  
  }
  
  $correlation = ((($n*$sumPxy)-($sumx*$sumy)) / sqrt(($n*$sumxs - ($sumx*$sumx))*($n*$sumys - ($sumy*$sumy))));
  return $correlation;
}



$words = array();    

for($i=1;$i<=100;$i++) array_push($words,$i);

$num = count($words);  
$w =0;
//The total number of possible combinations  
$total = $num * $num;  

$temp = array();
$comb = array();

//I NEED TO REMOVE GROUPS OF 6 ELEMENTS
$remove = 6;

//Loop through each possible combination   

//PROBABLY HERE IS THE BUG
//I'M GETTINT TO MUCH SIMILAR COMBINATIONS

for ($i = 0; $i < $total; $i++) {   
        //For each combination check if each bit is set  
    for ($j = 0; $j < $num; $j++) {  
       //Is bit $j set in $i?  
        //if (pow(2, $j) & $i) echo $words[$j] . ' ';       
	if (pow(2, $j) & $i) {
	  $w++;
	  $temp[] = $words[$j];
	}
	if($w == $remove) {
       $comb[$i] = $temp;
 	   $w = 0;
	   $temp = array();
	}
    }  
    
}

//REMOVING DUPLICATE VALUES
foreach($comb as $p) {
  $b[] = array_unique($p);
}

foreach($b as $c) {
if(count($c) == $remove) {
    $final[] = $c;
} 
}
$master = array();
foreach($final as &$t) {
  sort($t);
}

function multi_unique($array) {
        foreach ($array as $k=>$na)
            $new[$k] = serialize($na);
        $uniq = array_unique($new);
        foreach($uniq as $k=>$ser)
            $new1[$k] = unserialize($ser);
        return ($new1);
}

$final = multi_unique($final);



//REAL VALUES

$a = array(796.70, 425.70,371.68,725.48,1116.75,1453.04,1047.59,1147.30,2588.53,1824.40,2141.22,1279.14,1652.35,1657.36,2392.31,1759.18,2193.11,1121.48,1187.00,2337.72,2843.54,2387.14,3115.58,3674.36,432.19,4211.44,2665.15,432.19,1101.49,1108.37,345.53,517.25,1298.37,848.52,273.84,808.50,1135.06,521.70,347.80,343.43,460.73,686.87,578.14,235.12,243.44,235.12,417.27,599.99,699.99,830.31,782.89,522.68,632.55,709.32,2160.60,2065.75,1786.44,2025.76,3190.90,2109.16,1887.02,1652.60,3598.93,664.99,673.63,4493.26,1422.96,1181.20,500.44,449.46,540.98,457.587,89.02,743.12,6249.16,2103.99,578.13,1003.62,1748.03,1512.25,1782.85,465.75,1110.40,2658.48,987.93,1808.22,2510.86,1561.90,2947.39,4017.31,479.66,2386.98,1028.28,1301.45);

$b = array(1428.42,3093.44,1212.77,1308.62,1134.24,2493.63,3045.74,2759.02,3223.01,2571.50,2621.11,2222.28,1337.04,1286.16,1474.78,1145.49,2215.41,4233.91,2714.28,2629.93,2657.05,3474.56,749.13,3809.06,4533.39,1309.04,4085.50,4128.00,1437.19,4005.25,4695.29,3190.39,847.59,2664.70,3748.92,1839.42,739.07,1469.14,809.29,1026.29,835.76,822.52,392.89,476.22,452.81,675.79,1205.43,912.29,968.86,729.22,589.34,976.47,989.93,1010.01,976.47,1864.97,815.64,879.64,2957.40,2464.01,3107.68,2311.18,6746.77,3179.73,4074.20,3439.41,411.20,1753.65,1041.48,1000.40,602.91,753.56,698.63,2630.29,3082.19,8243.15,5828.21,3309.90,9521.83,3685.61,1846.11,3221.18,5082.69,3232.88,1851.27,671.83,3647.81,698.63,3836.90,1342.54,1520.81,1222.14,2726.06,1205.56,794.65,2123.34,935.39,890.86,4162.40,1538.44);

//MINUMUM CORRELATION
$correl = -1;
$keys = array();


foreach($final as $cur) {
  $tb = $b;
  
  foreach($cur as $p) {
    unset($tb[$p]);
  }
  $correlation = simplecorrelation($a,$tb);
  if($correlation > $correl) {
    $correl = $correlation;
$keys = $cur;
   }  
}



print $correl;

//KEYS WITCH REMOVED IMPROVE THE CORRELATIONS
print_r($keys);