ML.NET回归训练模型不影响预测

Question

我正在使用 ML.NET 回归模型来预测时间序列数据，是的，我知道有专门针对时间序列的类，但我需要它在我的论文中进行比较。但是当我更改模型甚至删除模型时，预测值不会改变。 这是带有模型的课程：

internal class FastTree
{
    internal static List<SingleValueForecast> predict(List<Value> trainData, int scope)
    {
        var context = new MLContext();
        var data = context.Data.LoadFromEnumerable(trainData);
        var pipeline = context.Transforms.CopyColumns("Label", "zatizeniCerpani")
            .Append(context.Transforms.Concatenate("Features", "Year", "Month", "Day","Hour"))
            .Append(context.Regression.Trainers.FastTree("Label", "Features"));
        var model = pipeline.Fit(data);
        var predictionEngine = context.Model.CreatePredictionEngine<Value, SingleValueForecast>(model);
        var forecasts = new List<SingleValueForecast>();
        var last = trainData.Last();
        forecasts.Add( new SingleValueForecast(last.date,last.zatizeniCerpani));
        for (int i = 0; i < scope; i++)
        {
            var input = forecasts.Last();

            var forecast = predictionEngine.Predict(input.toValue());
            forecasts.Add(forecast);
        }

        return forecasts;
    }

}

这是相关的类：

public class Value
{
    public DateTime date { get; set; }
    public float Year { get; set; }
    public float Month { get; set; }
    public float Day { get; set; }
    public float Hour { get; set; }
    public float zatizeniCerpani { get; set; }
    public Value(DateTime date, float zatizeniCerpani)
    {
        this.date = date;
        this.zatizeniCerpani = zatizeniCerpani;
        this.Year = date.Year;
        this.Month = date.Month;
        this.Day = date.Day;
        this.Hour = date.Hour;
    }
}

public class SingleValueForecast
{
    public float Year { get; set; }
    public float Month { get; set; }
    public float Day { get; set; }
    public float Hour { get; set; }
    public float zatizeniCerpani { get; set; }
    public DateTime date { get; set; }

    public SingleValueForecast(DateTime date, float zatizeniCerpani)
    {
        this.date = date;
        this.zatizeniCerpani = zatizeniCerpani;
        this.Year = date.Year;
        this.Month = date.Month;
        this.Day = date.Day;
        this.Hour = date.Hour;
    }

    public SingleValueForecast()
    { }

    public Value toValue()
    {
        return new Value(date,zatizeniCerpani);
    }

}

我期望在更改我正在使用的模型时值会发生变化。但对于任何模型，甚至没有模型，值都是相同的。

我将非常感谢您的帮助。

编辑： 好吧，我错过了很多东西，但主要是它没有输出列，所以我总是得到相同的值，这里是固定的 SingleValueForecast 类：

public class SingleValueForecast
{
    public float Year { get; set; }
    public float Month { get; set; }
    public float Day { get; set; }
    public float Hour { get; set; }
    public float zatizeniCerpani { get; set; }
    public DateTime date { get; set; }

    [ColumnName("Score")]
    public float PredictedZatizeniCerpani { get; set; }

    public SingleValueForecast(DateTime date, float zatizeniCerpani)
    {
        this.date = date;
        this.zatizeniCerpani = zatizeniCerpani;
        this.Year = date.Year;
        this.Month = date.Month;
        this.Day = date.Day;
        this.Hour = date.Hour;
    }

    public SingleValueForecast()
    { }

    public Value toValue()
    {
        return new Value(date,zatizeniCerpani);
    }

}

Answer 1

找到解决方案，这是我的错误。编辑了问题。