Blatt 8 Apriori Alg

blatt8/src/blatt8/Apriori.java

@@ -0,0 +1,205 @@
+package blatt8;
+
+import java.io.BufferedReader;
+import java.io.FileInputStream;
+import java.io.FileReader;
+import java.io.IOException;
+import java.io.InputStreamReader;
+import java.util.ArrayList;
+import java.util.HashMap;
+import java.util.List;
+import java.util.Map;
+import java.util.Random;
+
+public class Apriori {
+
+	
+	static List<int[]> itemsets = new ArrayList<int[]>();
+	static float minSup = 0.01f;
+	static String path = "transactions.txt";
+	static int numItems=0;
+	static int numTransactions=0;
+	
+	private static void analyzeTransactionFile(){
+		try (BufferedReader br = new BufferedReader(new FileReader(path))) {
+		    String line;
+		    while ((line = br.readLine()) != null) {
+		    	numTransactions++;
+		    	String[] products = line.split("\\s+");
+		    	for(int i=0;i<products.length;i++){
+		    		if(Integer.parseInt(products[i])+1>numItems){
+		    			numItems = Integer.parseInt(products[i])+1;
+		    		}
+		    	}
+		    }
+		}
+		catch(IOException e){
+			e.printStackTrace();
+		}
+		System.out.println("Num Transactions:"+numTransactions);
+		System.out.println("Num Items:"+numItems);
+	}
+	
+	
+	private static void AprioriAlgorithm(){
+		int k=1;
+		//Calculate L1
+		System.out.println(k+"-Itemsets");
+		find_frequent_1_itemsets();
+        for (k=2;itemsets.size()>0;k++)
+        {
+    		System.out.println("\n"+k+"-Itemsets");
+            generateCandidates();
+            
+            calculateFrequentItemsets();
+        } 
+	}
+	
+
+	private static void generateCandidates() {
+		int k = itemsets.get(0).length+1;
+		List<int[]> candidates = new ArrayList<int[]>();
+		for(int i=0; i<itemsets.size(); i++)
+        {
+            for(int j=0; j<itemsets.size(); j++)
+            {
+                int[] I1 = itemsets.get(i);
+                int[] I2 = itemsets.get(j);
+                boolean match=true;
+                for(int l=0;l<k-2;l++){
+                	if(I1[l]!=I2[l]){
+                		match = false;
+                		continue;
+                	}
+                }
+                if(match){
+                	if(I1[k-2]<I2[k-2]){
+                		int[] c = new int[k];
+                		for(int l=0;l<k-1;l++){
+                			c[l]=I1[l];
+                    	}
+                		c[k-1]=I2[k-2];
+                		if(!prune(c)){
+                			candidates.add(c);
+                		}
+                	}
+                }
+            }
+        }
+		itemsets=new ArrayList<int[]>(candidates);
+		if(itemsets.size()>0)System.out.println(itemsets.size()+" Candidates for k="+itemsets.get(0).length);
+		else System.out.println("0 Candidates");
+	}
+
+
+	private static boolean prune(int[] c) {
+//		if(new Random().nextInt(50)==0)return false;
+		return false;
+//		for(int skip=0;skip<c.length;skip++){
+//			int[] subset = new int[c.length-1];
+//			for(int i=0;i<c.length-1;i++){
+//				if(i>=skip)
+//					subset[i]=c[i+1];
+//				else
+//					subset[i]=c[i];
+//			}
+//			boolean match=false;
+//			for(int i=0;i<itemsets.size();i++){
+//				if(compare(itemsets.get(i),subset)){
+//					match=true;
+//					break;
+//				}
+//			}
+//			if(!match)
+//				return true;
+//			
+//		}
+//		return false;
+		
+	}
+	
+	private static boolean contains(int[] a, int[] b){
+		for(int j=0;j<b.length;j++){
+			boolean match=false;
+			for(int i=0;i<a.length;i++){
+				if(b[j]==a[i]){
+					match = true;
+				}
+			}
+			if(!match)return false;
+		}
+		return true;
+	}
+
+	private static boolean compare(int[] a, int[] b){
+		for(int i=0;i<a.length;i++){
+			if(a[i]!=b[i]){
+				return false;
+			}
+		}
+		return true;
+	}
+
+	private static void find_frequent_1_itemsets() {
+		itemsets = new ArrayList<int[]>();
+        for(int i=0; i<numItems; i++)
+        {
+        	int[] cand = {i};
+        	itemsets.add(cand);
+        }
+		System.out.println(itemsets.size()+" Candidates for k="+itemsets.get(0).length);
+        calculateFrequentItemsets();
+	}
+
+	private static void calculateFrequentItemsets()
+    {
+		List<int[]> frequent = new ArrayList<int[]>();
+		int[] counter = new int[itemsets.size()];
+		try (BufferedReader br = new BufferedReader(new FileReader(path))) {
+		    String line;
+		    while ((line = br.readLine()) != null) {
+		    	String[] products = line.split("\\s+");
+		    	int trans[] = new int[products.length];
+		    	for(int i=0;i<trans.length;i++){
+		    		trans[i]=Integer.parseInt(products[i]);
+		    	}
+		    	for(int cand=0;cand<itemsets.size();cand++){
+		    		if(contains(trans,itemsets.get(cand))){
+		    			counter[cand]++;
+		    		}
+		    	}
+		    }
+		}
+		catch(IOException e){
+			e.printStackTrace();
+		}
+		for(int i=0;i<counter.length;i++){
+			
+			if((float)counter[i]/(float)numTransactions>=minSup){
+				frequent.add(itemsets.get(i));
+				printItemset(itemsets.get(i));
+			}
+		}
+		itemsets = frequent;
+		if(itemsets.size()>0)System.out.println("Total Number Frequent "+itemsets.get(0).length+"-Itemsets: "+itemsets.size());
+		else System.out.println("0 Frequent");
+		
+    }
+
+	private static void printItemset(int[] is) {
+		String print="";
+		for(int i=0;i<is.length;i++){
+			print+=is[i]+" - ";
+		}
+		System.out.println(print);
+	}
+
+
+	public static void main(String[] args) {
+		analyzeTransactionFile();
+		AprioriAlgorithm();
+		
+		
+	}
+
+}