public simi[,] OrderedSimilarityMatrix = new simi[itemTotal, itemTotal]; //排序后的相似性矩阵
/// <summary>
/// 将数据集分成训练集 和 测试集 ----------------Success
/// </summary>
/// <param name="i"></param>
/// <param name="j"></param>
public void SplitDataSet(int i, int j)
{
Array.Clear(TestSet, 0, 3952 * 6040);
Array.Clear(TrainSet, 0, 3952 * 6040);
Thread.Sleep(1000);
string sqlStr = "select * from Ratings";
var dataReader = SqlSeverProvider.ExecuteReader(sqlStr);
while (dataReader.Read())
{
var userID = dataReader.GetInt32(0);
var itemID = dataReader.GetInt32(1);
var rate = dataReader.GetInt32(2);
Random rand = new Random(DateTime.Now.Millisecond);
try
{
if (rand.Next() % (i - 1) == j)
{
TestSet[itemID - 1, userID] = 1;
}
else
{
TrainSet[itemID - 1, userID - 1] = 1;
}
}
catch { }
}
}
//public int[] GetRecomend(int userId,List<int> ?itemIds)
//{
// List<int> items = new List<int>();
// items.Sort();
// return null;
//}
public int[] GetRecommend(int userId, params int[] itemIds)
{
Dictionary <int, double> recommendedList = new Dictionary <int, double>();
//根据用户id 获取最佳推荐的10个，添加到recommendedList
string sqlStr = string.Format("select top(10) itemID from UserInterest where userID={0} and interest !=0 order by interest desc", userId);
SqlSeverProvider.ExecuteReader(sqlStr);
//根据itemid 获取相似度较高的物品，添加到recommendedList
// select top(20)* from SimilarityMatrix where itemOne=176 order by similarity desc
// select top(10) itemID,COUNT(itemID) as total from Ratings group by itemID order by total desc 高访问率
//根据相似度高低对 recommendedList进行排序
var result = from it in recommendedList orderby it.Value select it;
foreach (var item in result.Take(10))
{
Console.WriteLine(item.Key);
}
return(null);
}