final_recommendor.php

<?php 
ini_set('max_execution_time', 300); //5 minuts

// PHP implementation of Pearson's product-moment correlation coefficient
function PCC($a, $b) {
    $mean_a = array_sum($a)/count($a);
	$mean_b = array_sum($b)/count($b);
	$sum_ab = 0;
    $sum_a = 0;
    $sum_b = 0;
    $sum_a_sqr = 0;
    $sum_b_sqr = 0;
    //$n = min(array(count($a), count($b)));
    for ($i = 0; $i < count($a); $i++) {
        if (!isset($a[$i]) || !isset($b[$i])) { continue; }
        $sum_ab += ($a[$i]-$mean_a) * ($b[$i]-$mean_b);
        $sum_a += $a[$i]-$mean_a;
        $sum_b += $b[$i]-$mean_b;
        $sum_a_sqr += pow($a[$i]-$mean_a, 2);
        $sum_b_sqr += pow($b[$i]-$mean_b, 2);
    }
	if((sqrt($sum_a_sqr) * sqrt($sum_b_sqr))!=0){
		return $sum_ab / (sqrt($sum_a_sqr) * sqrt($sum_b_sqr));		
	}else{
		return -1;// not similar
	}
	
}
/* 
$a1 = array(0, 1, 1, 0, 1, 0, 0, 1, 1, 0, 0, 0, 1, 0);
$b1 = array(1, 0, 0, 1, 1, 1, 1, 0, 0, 1, 1, 1, 0, 1);
echo corr($a1, $b1) . "\n<br/>";

$a2 = array(1, 1, 1, 1, 0, 0, 1, 1, 1, 0, 0, 1, 0, 1);
$b2 = array(1, 1, 0, 1, 0, 1, 1, 0, 1, 1, 0, 1, 0, 0);
echo corr($a2, $b2) . "\n<br/>";

$a3 = array(1, 0, 0, 0, 1, 0, 0, 1, 1, 0, 0, 0, 1, 0);
$b3 = array(1, 0, 0, 0, 1, 0, 0, 1, 1, 0, 0, 0, 1, 0);
echo corr($a3, $b3) . "\n<br/>"; */
require_once("db/dbcon.php");
function getUsers(){
	$dbCon = new databaseManager();
	$query="select distinct(userID) as userID from rating where ratingValue is not null";
	$userData = $dbCon->executeQuery($query, array(0), "sread");
	$users = array();
	foreach($userData as $userKey => $userValue){
		$users[] = $userValue['userID'];
	}
	return $users;
}
function getUnratedProductsOfUser($userID, $itemID){
	$dbCon = new databaseManager();
	$query="select distinct(userID) as user from rating where itemID=? and ratingValue is null";
	$userData = $dbCon->executeQuery($query, array($userID, $itemID), "cread");
	$p = array();
	foreach($userData as $userKey => $userValue){
		$p[] = $userValue['item'];
	}
	return $p;
}
function predictToAllUsers(){
	$p = getRatedProduct();
	$pridictedProducts = array();
	for($i=0;$i<count($p);$i++){
		$users = getUnratedItems($p[$i]);
		for($j=0;$j<count($users);$j++){			
			echo "Predict product $p[$i] to user $users[$j] on the basis of: ";print_r(predictProduct($p[$i], $users[$j]));echo "<br>";
			$pridictedProducts[]=array($users[$j]=>$p[$i]);
		}
	}
	return $pridictedProducts;
}
/* function getProducts(){
	$dbCon = new databaseManager();
	$query="select distinct(itemID) as productID from rating where ratingValue is not null";
	$ratedProducts = $dbCon->executeQuery($query, array(0), "sread");
	$products = array();
	foreach($ratedProducts as $productKey => $productValue){
		$products[] = $productValue['productID'];
	}
	//var_dump($products);
	return $products;
} */

function getRatedProduct(){
	$dbCon = new databaseManager();
	$query="select distinct(itemID) as productID from rating where ratingValue is not null";
	$ratedProducts = $dbCon->executeQuery($query, array(0), "sread");
	$products = array();
	foreach($ratedProducts as $productKey => $productValue){
		$products[] = $productValue['productID'];
	}
	//var_dump($products);
	return $products;
}
function getRatedColumnValues($itemID){
	$dbCon = new databaseManager();
	$query="select ratingValue from rating where itemID=?";
	$data = $dbCon->executeQuery($query, array($itemID), "cread");
	$a =array();
	foreach($data as $key=>$value){
		$a[] = $value['ratingValue'];
	}
	return $a;
}
function getUnratedItems($itemID){
	$unrated=array();
	$dbCon = new databaseManager();
	$query="select userID from rating where itemID=? and ratingValue is null";
	$data = $dbCon->executeQuery($query, array($itemID), "cread");
	foreach($data as $key=>$value){
		$unrated[] = $value['userID'];
	}
	return $unrated;
}

//$users= getUsers();\

function getSimilarValues(){
	$products = getRatedProduct();
	$simValues = array(array());
	for($j=0;$j<sizeOf($products);$j++){
		$valueOfj = getRatedColumnValues($products[$j]);
		for($k = 0; $k <sizeOf($products); $k++){
			$valueOfk = getRatedColumnValues($products[$k]);
			//echo "similarity of product ".$products[$j]." with ".$products[$k]." is ".PCC($valueOfj,$valueOfk)."\n<br/>";
			$simValues[$j][$k] = PCC($valueOfj,$valueOfk);
		}
	}
	return $simValues;
}
function displaySimilarValues($simValues){
	echo "<h1>Similarity Values Table</h1><hr>";
	echo "<table border='1'>";
	for($i=0; $i<sizeOf($simValues); $i++){
		echo "<tr>";
		for($j=0;$j<sizeOf($simValues); $j++){
			echo "<td>".$simValues[$i][$j]."</td>";
		}
		echo "</tr>";
	}
	echo "</table >";
}


function getSimilarityOf($itemID){
	$products = getRatedProduct();
	$simValues = array();
	for($j=0;$j<sizeOf($products);$j++){
		for($k = 0; $k <sizeOf($products); $k++){
			if($products[$k] == $itemID){
				$valueOfj = getRatedColumnValues($products[$j]);
				$valueOfk = getRatedColumnValues($products[$k]);
				$simValues[$j] = PCC($valueOfj,$valueOfk);				
			}
		}
	}
	return $simValues;
}

#$users = getUsers();
function predictProduct($productID, $toUser){
	$dbCon = new databaseManager();
	$query="select itemID, ratingValue from rating where userID=? and ratingValue is not null";
	$data = $dbCon->executeQuery($query, array($toUser), "cread");
	/* $sum_sim_r = 0;
	$sum_sim = 0; */
	$threshold=0.2;//75%
	$rv =array();
	$similarProducts =array();
	foreach($data as $key=>$value){
		//echo "ItemID". $value['itemID']."====>".$value['ratingValue']."<br>";
		$rv[]=$value['ratingValue'];
	}
	foreach($data as $key=>$value){
		//echo "ItemID". $value['itemID']."====>".$value['ratingValue']."<br>";
		
		//$sum_sim_r += PCC(getRatedColumnValues($productID), $rv)*$value['ratingValue'];
		//$sum_sim += PCC(getRatedColumnValues($productID), $rv);
		//print_r(PCC($rv,getRatedColumnValues($productID)));
		if(PCC($rv,getRatedColumnValues($productID)) > $threshold){
			$similarProducts[] = $value['itemID'];
		}
		
	}
	/* if($sum_sim !=0){
		return $sum_sim_r/$sum_sim;
	}else{
		return 0;
	} */
	return $similarProducts;
}	

echo "Rating Prediction of item: 84 for user: 52 is:  ";print_r(predictProduct(15, 51));echo "<br>";
//echo "Rating Prediction of item: 84 for user: 52 is:  ";print_r(predictProduct(8, 53));echo "<br>";
//echo "Rating Prediction of item: 84 for user: 52 is:  ";print_r(predictProduct(8, 54));echo "<br>";
$ratedItem = getRatedProduct();
echo "<h1>Rated Product ID's</h1>";
foreach($ratedItem as $item){
	echo $item."\n<br>";
}
echo "<hr>";
echo "<h1>Un-rated Product ID's</h1>";
$a =getUnratedItems(7);
for($i=0;$i<sizeOf($a);$i++){
	echo $a[$i]."\n<br>";
}
echo "<hr>";

displaySimilarValues(getSimilarValues());
/* echo "<pre>";
print_r(predictToAllUsers());
echo "</pre>"; */

function predictToSingleUser($userID){
	$userItems=array();
	$data = predictToAllUsers();
	for($i=0;$i<count($data);$i++){
		foreach($data[$i] as $key=>$value){
			if($key==$userID){
				echo $userItems[] = $data[$i][$key]."<br>";
			}
		}
	}
	return $userItems;
}
print_r(predictToSingleUser(65));
?>