Apriori算法Java实现示例 - JAVA

package xx;

import java.util.HashMap;
import java.util.HashSet;
import java.util.Iterator;
import java.util.LinkedList;
import java.util.List;
import java.util.Map;
import java.util.Map.Entry;
import java.util.Set;
import java.util.TreeMap;
import java.util.TreeSet;

public class Apriori {

	public static void main(String[] args) throws Exception {

		// 初始化事务集
		List> trans = new LinkedList>();
		trans.add(new ItemSet(new String[] { "I1", "I2", "I5" }));
		trans.add(new ItemSet(new String[] { "I2", "I4" }));
		trans.add(new ItemSet(new String[] { "I2", "I3" }));
		trans.add(new ItemSet(new String[] { "I1", "I2", "I4" }));
		trans.add(new ItemSet(new String[] { "I1", "I3" }));
		trans.add(new ItemSet(new String[] { "I2", "I3" }));
		trans.add(new ItemSet(new String[] { "I1", "I3" }));
		trans.add(new ItemSet(new String[] { "I1", "I2", "I3", "I5" }));
		trans.add(new ItemSet(new String[] { "I1", "I2", "I3" }));

		int MSF = 2; // 设定最小支持频次为2

		Map> rst = findFrequentItemSets(trans, MSF);

		// 输出频繁项集
		System.out.println("Frequent Item Sets:");
		for (Entry> entry : rst.entrySet()) {
			Integer itemSetSize = entry.getKey();
			System.out.printf("Frequent %d Item Sets:\n", itemSetSize);
			for (ItemSet set : entry.getValue())
				System.out.printf("%s, %d\n", set, set.frequence);
		}

		double MCONF = 0.6; // 设定最小置信度为60%

		Map directMap = new HashMap();
		for (Entry> entry : rst.entrySet()) {
			for (ItemSet set : entry.getValue())
				directMap.put(set, set);
		}

		// 根据频繁项集构造关联规则
		System.out.println();
		System.out.println("Association Rules:");
		for (Entry> entry : rst.entrySet()) {
			for (ItemSet set : entry.getValue()) {
				double cnt1 = directMap.get(set).frequence;
				List




 subSets = set.listNotEmptySubItemSets();
				for (ItemSet subSet : subSets) {
					int cnt2 = directMap.get(subSet).frequence;
					double conf = cnt1 / cnt2;
					if (cnt1 / cnt2 >= MCONF) {
						ItemSet remainSet = new ItemSet();
						remainSet.addAll(set);
						remainSet.removeAll(subSet);
						System.out.printf("%s => %s, %.2f\n", subSet,
								remainSet, conf);
					}
				}
			}
		}
	}

	/**
	 * 查找事务集中的所有频繁项集，返回Map为：L -> 所有频繁L项集的列表
	 */
	static Map> findFrequentItemSets(
			Iterable> transIterable, int MSF) {
		Map> ret = new TreeMap>();

		// 首先确定频繁1项集
		Iterator> it = transIterable.iterator();
		Set oneItemSets = findFrequentOneItemSets(it, MSF);
		ret.put(1, oneItemSets);

		int preItemSetSize = 1;
		Set preItemSets = oneItemSets;

		// 基于获得的所有频繁L-1项集迭代查找所有频繁L项集，直到不存在频繁L-1项集
		while (!preItemSets.isEmpty()) {
			int curItemSetSize = preItemSetSize + 1;

			// 获取频繁L项集的所有候选L项集
			List candidates = aprioriGenCandidates(preItemSets);

			// 扫描事务集以确定所有候选L项集出现的频次
			it = transIterable.iterator();
			while (it.hasNext()) {
				Set tran = it.next();
				for (ItemSet candidate : candidates)
					if (tran.containsAll(candidate))
						candidate.frequence++;
			}

			// 将出现频次不小于最小支持频次的候选L项集选为频繁L项集
			Set curItemSets = new HashSet();
			for (ItemSet candidate : candidates)
				if (candidate.frequence >= MSF)
					curItemSets.add(candidate);
			if (!curItemSets.isEmpty())
				ret.put(curItemSetSize
Apriori算法Java实现示例(一)
