static DynamicKMeans TrainModel(List <Song> songs, int clustersCount)
{
DynamicKMeans kMeans = new DynamicKMeans(songs, clustersCount);
kMeans.Clusterize();
// ConsoleHelper.ShowClusters(kMeans);
return(kMeans);
}
public static void Main(string[] args)
{
AppConfig appConfig = FileSystemHelper.GetAppConfig();
DateTime runDate = GetRunDate();
List <Song> songs = GetPreparedTrainingData();
int experimentIndex = 0;
foreach (ExperimentConfig config in appConfig.ExperimentConfigs)
{
for (int i = 0; i < config.ModelsCount; i++)
{
DynamicKMeans trainedKMeans = TrainModel(songs, config.ClustersCount);
FileSystemHelper.SaveKMeansModel(trainedKMeans.Model, runDate, experimentIndex);
experimentIndex += 1;
}
}
List <StaticKMeans> models = ReadAllModels(songs);
List <ModelAnalyzer> analyzers = models.Select(model =>
{
ModelAnalyzer analyzer = new ModelAnalyzer(model);
analyzer.Analyze();
return(analyzer);
}).ToList();
AnalysisDrawer drawer = new AnalysisDrawer(analyzers);
drawer.CreatePlots();
FileSystemHelper.SaveAnalysis(drawer, runDate);
List <Song> analyzingSongs = GetAnalyzingSongs();
FileSystemHelper.ClearPredictions();
for (int i = 0; i < models.Count; i++)
{
StaticKMeans model = models[i];
List <Prediction> predictions = MakePredictions(model, analyzingSongs);
FileSystemHelper.SavePredictions(predictions, runDate, i);
}
}