public async Task <int> CreatePredictionCourseForParticipantAsync(RatingCursoEntity ratingCursoEntity)
{
if (_dbConnection.State == ConnectionState.Closed)
{
_dbConnection.Open();
}
var result = await _dbConnection.QueryAsync <int>("CreatePredictionCurso", new { ratingCursoEntity.Rating, ratingCursoEntity.IdParticipante, ratingCursoEntity.IdCurso },
commandType : CommandType.StoredProcedure);
return(result.FirstOrDefault());
}
public async Task <bool> CreatePredictionCoursesForParticipants()
{
Stopwatch stopWatch = new Stopwatch();
stopWatch.Start();
Log.Information("Inicio CreatePredictionCoursesForParticipants");
try
{
var ratingCursos = await ratingCursoRepository.GetAllAsync();
var ratingCursosToTrain = ratingCursos.Select(x => new CourseEntry {
IdCurso = Convert.ToUInt32(x.IdCurso), IdParticipante = Convert.ToUInt32(x.IdParticipante), Rating = (float)x.Rating
}).ToList();
MLContext mlContext = new MLContext();
var traindata = mlContext.Data.LoadFromEnumerable(ratingCursosToTrain);
MatrixFactorizationTrainer.Options options = new MatrixFactorizationTrainer.Options();
options.MatrixColumnIndexColumnName = nameof(CourseEntry.IdCurso);
options.MatrixRowIndexColumnName = nameof(CourseEntry.IdParticipante);
options.LabelColumnName = "Rating";
options.LossFunction = MatrixFactorizationTrainer.LossFunctionType.SquareLossRegression;
options.Alpha = 0.01;
options.Lambda = 0.025;
options.NumberOfIterations = 1000;
var trainingPipeLine = mlContext.Recommendation().Trainers.MatrixFactorization(options);
ITransformer model = trainingPipeLine.Fit(traindata);
var predictionengine = mlContext.Model.CreatePredictionEngine <CourseEntry, CursoRatingPrediction>(model);
int?allItemsInteger = default;
var cursos = await cursoRepository.GetCursosAsync(allItemsInteger, allItemsInteger, allItemsInteger, default);
var participantes = await participanteRepository.GetParticipantesAsync(allItemsInteger, allItemsInteger);
await predictionTrainerRepository.ClearPredictionsAsync();
Log.Information("All predictions has deleted until {date}", DateTime.Now);
foreach (var participante in participantes)
{
var topCursos = cursos.Select(x => new
{
Curso = x.Nombre,
x.IdCurso,
Score = (predictionengine.Predict(
new CourseEntry()
{
IdCurso = Convert.ToUInt16(x.IdCurso),
IdParticipante = Convert.ToUInt16(participante.IdParticipante)
}).Score)
}).OrderByDescending(x => x.Score).Take(5);
foreach (var item in topCursos)
{
Console.WriteLine($" Participante: {participante.IdParticipante} {participante.Nombre} Curso: {item.Curso} prediction.Rating: {Math.Round(item.Score)} {item.Score} ");
RatingCursoEntity ratingCursoEntity = new RatingCursoEntity();
var rating = (int)Math.Round(item.Score);
ratingCursoEntity.Rating = rating >= (int)decimal.Zero ? rating : (int)decimal.Zero;
ratingCursoEntity.IdParticipante = participante.IdParticipante;
ratingCursoEntity.IdCurso = item.IdCurso;
await predictionTrainerRepository.CreatePredictionCourseForParticipantAsync(ratingCursoEntity);
}
}
stopWatch.Stop();
Log.Information("Finalizado CreatePredictionCoursesForParticipants in {elapsedTime}", stopWatch.Elapsed);
return(true);
}
catch (Exception ex)
{
stopWatch.Stop();
Log.Error("Un error a ocurrido en la Predicción de Cursos con la siguiente excepción: {errormessage}", ex.Message);
return(false);
}
}
