public SvdImpl(byte[,] ratings, IdMap userIds, IdMap itemIds)
{
_ratings = ratings;
_userIds = userIds;
_itemIds = itemIds;
_uCount = ratings.GetLength(0);
_iCount = ratings.GetLength(1);
_baseUser = new double[_uCount];
_baseItem = new double[_iCount];
_p = new double[_uCount, Size];
_q = new double[_iCount, Size];
Random prng = new Random();
for (int i = 0; i < _uCount; i++)
{
for (int j = 0; j < Size; j++)
{
_p[i, j] = prng.NextDouble();
}
}
for (int i = 0; i < _iCount; i++)
{
for (int j = 0; j < Size; j++)
{
_q[i, j] = prng.NextDouble();
}
}
}
private static void Main(string[] args)
{
if (args.Length < 3)
{
Console.WriteLine("too few arguments");
return;
}
string testFile = args[0];
string resFile = args[1];
IdMap userIds = new IdMap(3500);
IdMap itemIds = new IdMap(5000);
List <UserRatingModel> models = new List <UserRatingModel>(1000000);
for (int i = 2; i < args.Length; i++)
{
ReadTrainingFile(args[i], models, userIds, itemIds);
}
byte[,] ratings = new byte[userIds.Count, itemIds.Count];
foreach (var model in models)
{
ratings[model.UserId, model.ItemId] = model.Rating;
}
SvdImpl svd = new SvdImpl(ratings, userIds, itemIds);
svd.Learn();
using (var reader = new StreamReader(File.OpenRead(testFile)))
using (var writer = new StreamWriter(File.OpenWrite(resFile)))
{
reader.ReadLine();
writer.WriteLine("id,rating");
string line;
while ((line = reader.ReadLine()) != null)
{
string[] split = line.Split(',');
int rating = svd.PredictRating(long.Parse(split[1]), long.Parse(split[2]));
writer.WriteLine("{0},{1}", split[0], rating);
}
}
}
private static void ReadTrainingFile(string trainFile, List <UserRatingModel> models, IdMap userIds, IdMap itemIds)
{
using (var reader = new StreamReader(File.OpenRead(trainFile)))
{
reader.ReadLine();
string line;
while ((line = reader.ReadLine()) != null)
{
string[] split = line.Split(',');
models.Add(new UserRatingModel
{
UserId = userIds.GetOrInsert(long.Parse(split[0])),
ItemId = itemIds.GetOrInsert(long.Parse(split[1])),
Rating = byte.Parse(split[2])
});
}
}
}