基于社交网络的用户与基于物品的协同过滤推荐算法-java
完整工程+数据源：https://github.com/scnuxiaotao/recom_sys
import java.io.BufferedReader;
import java.io.FileNotFoundException;
import java.io.FileReader;
import java.io.FileWriter;
import java.io.IOException;
import java.util.HashMap;
public class itemcf {
/*
*
* 主函数~
*
*/
/*public static void main(String[] args) throws IOException {
_Run();
}*/
/*
*
* 基于物品的实现~
*
*/
static int usersum = 20836; //用户数
static int itemsum = 200; //物品总数
static int N = 3; //推荐个数
static int[][] train; //训练集合user item rate矩阵
static int[][] test;//测试集合user item rate矩阵
static double[][] trainuseritem; //训练集合user item 兴趣程度 矩阵
static int[][] recommend; //为每个用户推荐N个物品
static simi [][]simiItem; //排序后的相似性矩阵
static double [][]itemsim; //未排序的相似性矩阵
static String road = "data/6 总评论情感分析结果/酒店-评论（已转化）.txt"; //数据路径，格式为用户编号::物品编号::评分
static String road2 = "data/10 推荐/(已转化)天河酒店ID.txt";
static String road3 = "data/10 推荐/物品推荐.txt";
public static String road4 = "data/10 推荐/(已转化)用户ID.txt";
public static class simi
{
double value; //相似值
int num; //相似物品号
};
public static void _Run() throws IOException {
get_user_hotel_num();
System.out.println("usersum: "+usersum);
System.out.println("itemsum: "+itemsum);
train = new int[itemsum][usersum]; train[0][0] = 0; //训练集合user item rate矩阵
test = new int[itemsum][usersum]; test[0][0] = 0; //测试集合user item rate矩阵
trainuseritem =
new double[usersum][itemsum]; trainuseritem[0][0] = 0.0; //训练集合user item 兴趣程度 矩阵
recommend = new int[usersum][N]; recommend[0][0] = 0; //为每个用户推荐N个物品
simiItem = new simi[itemsum][itemsum]; //排序后的相似性矩阵
itemsim = new double[itemsum][itemsum]; //未排序的相似性矩阵
int i,j,k = 8; //去用户的k个最近邻居（相似度最高）来计算推荐物品
for(i = 0 ;i < itemsum;++i)
for(j = 0 ;j < itemsum;++j) simiItem[i][j] = new simi();
System.out.println("1.训练集");
SplitData(8,1);
//输出初始化的矩阵
/*for (i=0;i<10;i++)
{
System.out.println("Item"+i+": ");
for (j=0;j<5;j++)
{
System.out.print(train[i][j]+" ");
}
System.out.println();
}*/
System.out.println("2.计算物品之间相似性，得到相似性矩阵");
for (i=0;i0&&ItemB[i]>0)
{
comUser++;//查找ItemA与ItemB的都被用户评论的用户个数
}
if (ItemA[i]>0){
countIa++;//评论ItemA的用户数量
}
if (ItemB[i]>0){
countIb++;//评论ItemB的用户数量
}
}
double tem = Math.sqrt(countIa*countIb);
//double tem = 1;
//System.out.println(tem);
if(tem == 0)
{
return 0;
}
else
{
simility = comUser/tem;
return simility;
}
}
/*物品相似性矩阵排序（根据相似性由高到低排序）*/
public static void quickSort(int x, int start, int end) {
if (start < end) {
double base = simiItem[x][start].value; // 选定的基准值（第一个数值作为基准值）
double temp; // 记录临时中间值
int i_tmp;
int i = start, j = end;
do {
while ((simiItem[x][i].value > base) && (i < end))
i++;
while ((simiItem[x][j].value < base) && (j > start))
j--;
if (i <= j) {
temp = simiItem[x][i].value;
simiItem[x][i].value = simiItem[x][j].value;
simiItem[x][j].value = temp;
i_tmp = simiItem[x][i].num;
simiItem[x][i].num = simiItem[x][j].num;
simiItem[x][j].num = i_tmp;
i++;
j--;
}
} while (i <= j);
if (start < j)
quickSort(x, start, j);
if (end > i)
quickSort(x, i, end);
}
}
public static int sort()
{
for (int i=0;i0)//若这个用户同样对相似物品也有过行为
{
trainuseritem[i][j]+=simiItem[j][x].value;
}
}
return trainuseritem[i][j];
}
/*通过物品兴趣程度，推荐前N个*/
public static int getRecommend() //有bug,已修改
{
int maxnum;//当前最感兴趣物品号
for(int i=0;i package WjPack;
import java.io.BufferedReader;
import java.io.FileNotFoundException;
import java.io.FileReader;
import java.io.FileWriter;
import java.io.IOException;
import java.util.ArrayList;
import java.util.HashMap;
import java.util.HashSet;
import java.util.Iterator;
import java.util.Map;
import java.util.Set;
public class new_ojld_dis {
/*
*
* 主函数~
*
*/
/*public static void main(String[] args) throws IOException {
run();
}*/
/*
*
* 基于用户的实现~
*
*/
static String road_main = "data";
static String road = road_main + "/6 总评论情感分析结果/酒店-评论（已转化）.txt";//数据路径，格式为用户编号::物品编号::评分
static String road2 = road_main + "/10 推荐/(已转化)用户-关注.txt"; //数据路径，格式为用户编号::关注编号
static String road3 = road_main + "/10 推荐/(已转化)天河酒店ID.txt";
static String road4 = road_main + "/10 推荐/用户推荐.txt";
static String road5 = road_main + "/10 推荐/(已转化)用户ID.txt";
static int usersum = 20836; //用户数
static int itemsum = 200; //物品总数
static Map> score = new HashMap>();
static Set userSet = new HashSet();
static Set filmSet = new HashSet();
static FileWriter txtw;
static String tjhotel = "";
static ArrayList arr;
static {
arr = new ArrayList();
try {
score = get_score_from_road();
} catch (IOException e) {
}
}
public static void run() throws IOException {
txtw=new FileWriter(road4);
get_hotelid(road3);
for(int m = 0;m < 20836;++m) {
tjhotel = "";
new_ojld_dis.outNearbyUserList(arr.get(m));
if(tjhotel.length() > 1)
txtw.write("user"+arr.get(m)+tjhotel+"\r\n");
}
txtw.close();
}
public static void init() {
try {
get_user_hotel_num();
} catch (IOException e1) {
// TODO Auto-generated catch block
e1.printStackTrace();
}
System.out.println(usersum);
System.out.println(itemsum);
}
public static void get_user_hotel_num() throws IOException {
FileReader data_about = new FileReader(road3);
BufferedReader read_data_about=new BufferedReader(data_about);
int num = 0;
while(read_data_about.readLine() != null) num++;
itemsum = num;
data_about.close();
read_data_about.close();
data_about = new FileReader(road5);
read_data_about=new BufferedReader(data_about);
num = 0;
while(read_data_about.readLine() != null) num++;
usersum = num;
data_about.close();
read_data_about.close();
}
public static Map> get_score_from_road() throws IOException {
init();
String []tmps = new String[5];
FileReader data_about=new FileReader(road);
BufferedReader read_data_about=new BufferedReader(data_about);
String s2; //暂存文件一行记录
String usertmp = null;
Map> score = new HashMap>();
HashMap tempScore = new HashMap();
while((s2=read_data_about.readLine())!=null){
//寻找数据集每条记录对应的用户号和物品号
tmps = s2.split("::");
String username = tmps[0];
String filmname = tmps[1];
Integer socrename = Integer.valueOf(tmps[2]);
if(usertmp == null) {usertmp = username;arr.add(usertmp);}
else if(!usertmp.equals(username)) {
score.put(usertmp, tempScore);
usertmp = username;
arr.add(usertmp);
tempScore = new HashMap();
}
tempScore.put(filmname, socrename);
}
score.put(usertmp, tempScore);
arr.add(usertmp);
return score;
}
public static void outNearbyUserList(String user) throws IOException {
FileReader data_about=new FileReader(road2);
BufferedReader read_data_about=new BufferedReader(data_about);
Map scores = new HashMap();
String []tmps = new String[5];
HashMap thing = new HashMap();
String id; //暂存文件一行记录
int num = 0;
while((id=read_data_about.readLine())!=null){
tmps = id.split("::");
String username = tmps[0];
String fansname = tmps[1];
if(username.equals(user)) {
thing.put(fansname, 1);
}
}
for (int m = 0;m < arr.size()-1;++m) {
String tempUser = arr.get(m);
if (tempUser.equals(user) || !thing.containsKey(tempUser)) {
continue;
}
double score = getOSScore(user, tempUser);
if(score >= 0)
scores.put(tempUser, score);
}
data_about.close();
read_data_about.close();
}
private static Double getOSScore(String user1, String user2) throws NumberFormatException, IOException {
HashMap user1Score = (HashMap) score.get(user1);
HashMap user2Score = (HashMap) score.get(user2);
double totalscore = 100;
ArrayList hobby = new ArrayList();
Iterator it = user1Score.keySet().iterator();
while (it.hasNext()) {
String film = (String) it.next();
int a1 = (Integer) user1Score.get(film);
//System.out.println(film);
if(user2Score.get(film) == null) continue;
int b1 = (Integer) user2Score.get(film);
int a = a1 * a1 - b1 * b1;
//System.out.println(Math.abs(a));
totalscore = Math.sqrt(Math.abs(a));
}
if(totalscore == 0) {
int ok = 0;
it = user2Score.keySet().iterator();
if(it != null) {
while (it.hasNext()) {
String film = (String) it.next();
if(user1Score.get(film) == null) {
if(ok == 0) {
ok = 1;
}
tjhotel+="::"+hotel[Integer.parseInt(film)];
}
}
}
}
return totalscore;
}
static String []hotel = new String[201];
public static void get_hotelid(String r1) throws IOException {
FileReader data_about=new FileReader(r1);
BufferedReader read_data_about=new BufferedReader(data_about);
String id; //暂存文件一行记录
String []tmps = new String[5];
while((id=read_data_about.readLine())!=null){
tmps = id.split("::");
String hotelname = tmps[0];
String number = tmps[1];
hotel[Integer.parseInt(number)] = hotelname;
}
data_about.close();
read_data_about.close();
}
} 课程设计写的代码，可以用的，不过不写注释，但也不是很难看懂，先了解以下原理再看看代码就差不多了～
PS：因为抓到的用户ID和酒店都是数据都是类似434132这么长的编号，为了方便我用数组存，事先我全部转化为1开始的编号了。比如说有两个4654654,32131321，那我就转成1,2了，处理完推荐完再把1,2转成4654654,32131321