static void Main(string[] args)
{
var context = new MLContext();
var data = context.Data.LoadFromTextFile <HousingData>("./housing.csv", hasHeader: true, separatorChar: ',');
var features = data.Schema
.Select(col => col.Name)
.Where(colName => colName != "Label" && colName != "OceanProximity")
.ToArray();
var pipeline = context.Transforms.Text.FeaturizeText("Text", "OceanProximity")
.Append(context.Transforms.Concatenate("Features", features))
.Append(context.Regression.Trainers.LbfgsPoissonRegression());
var model = pipeline.Fit(data);
var transformedData = model.Transform(data);
// Get weights of model
var linearModel = model.LastTransformer.Model;
var weights = linearModel.Weights;
Console.WriteLine("Weights:");
for (int i = 0; i < features.Length; i++)
{
Console.WriteLine($"Feature {features[i]} has weight {weights[i]}");
}
Console.WriteLine(Environment.NewLine);
// Get global feature importance
var lastTransformer = model.LastTransformer;
var featureImportance = context.Regression.PermutationFeatureImportance(lastTransformer, transformedData);
Console.WriteLine("Global feature importance:");
for (int i = 0; i < featureImportance.Count(); i++)
{
Console.WriteLine($"Feature - {features[i]}: Difference in RMS - {featureImportance[i].RootMeanSquaredError.Mean}");
}
Console.WriteLine(Environment.NewLine);
// Get feature importance for each row
var firstRow = model.Transform(context.Data.TakeRows(transformedData, 1));
var featureContribution = context.Transforms.CalculateFeatureContribution(lastTransformer, normalize: false);
var featureContributionTransformer = featureContribution.Fit(firstRow);
var featureContributionPipeline = model.Append(featureContributionTransformer);
var predictionEngine = context.Model.CreatePredictionEngine <HousingData, HousingPrediction>(featureContributionPipeline);
var sampleData = new HousingData
{
Longitude = -122.25f,
Latitude = 37.85f,
HousingMedianAge = 55.0f,
TotalRooms = 1627.0f,
TotalBedrooms = 235.0f,
Population = 322.0f,
Households = 120.0f,
MedianIncome = 8.3014f,
OceanProximity = "NEAR BAY"
};
var prediction = predictionEngine.Predict(sampleData);
Console.WriteLine("Row feature importance:");
for (int i = 0; i < prediction.FeatureContributions.Length; i++)
{
Console.WriteLine($"Feature {features[i]} has feature contribution of {prediction.FeatureContributions[i]}");
}
}
static void Main(string[] args)
{
var context = new MLContext();
var data = context.Data.LoadFromTextFile <HousingData>("./housing.csv", hasHeader: true, separatorChar: ',');
var featureColumns = data.Schema
.Select(col => col.Name)
.Where(colName => colName != "Label" && colName != "OceanProximity")
.ToArray();
var pipeline = context.Transforms.Text.FeaturizeText("Text", "OceanProximity")
.Append(context.Transforms.Concatenate("Features", featureColumns))
.Append(context.Regression.Trainers.LbfgsPoissonRegression());
var model = pipeline.Fit(data);
var transformedData = model.Transform(data);
// Get weights of model
var linearModel = model.LastTransformer.Model;
var weights = linearModel.Weights;
Console.WriteLine("Weights:");
// Order features by importance
var featureWeights = weights
.Select((item, index) => new { index, item })
.OrderByDescending(theWeight => Math.Abs(theWeight.item));
foreach (var featureWeight in featureWeights)
{
Console.WriteLine($"Feature {featureColumns[featureWeight.index]} has weight {featureWeight.item}");
}
Console.WriteLine(Environment.NewLine);
// Get global feature importance
var lastTransformer = model.LastTransformer;
var featureImportance = context.Regression.PermutationFeatureImportance(lastTransformer, transformedData);
// Order features by importance
var featureImportanceMetrics =
featureImportance
.Select((metric, index) => new { index, metric.RSquared })
.OrderByDescending(myFeatures => Math.Abs(myFeatures.RSquared.Mean));
Console.WriteLine("Feature\tPFI");
foreach (var feature in featureImportanceMetrics)
{
Console.WriteLine($"{featureColumns[feature.index],-20}|\t{feature.RSquared.Mean:F6}");
}
Console.WriteLine(Environment.NewLine);
// Get feature importance for each row
var firstRow = model.Transform(context.Data.TakeRows(transformedData, 1));
var featureContribution = context.Transforms.CalculateFeatureContribution(lastTransformer, normalize: false);
var featureContributionTransformer = featureContribution.Fit(firstRow);
var featureContributionPipeline = model.Append(featureContributionTransformer);
var predictionEngine = context.Model.CreatePredictionEngine <HousingData, HousingPrediction>(featureContributionPipeline);
var sampleData = new HousingData
{
Longitude = -122.25f,
Latitude = 37.85f,
HousingMedianAge = 55.0f,
TotalRooms = 1627.0f,
TotalBedrooms = 235.0f,
Population = 322.0f,
Households = 120.0f,
MedianIncome = 8.3014f,
OceanProximity = "NEAR BAY"
};
var prediction = predictionEngine.Predict(sampleData);
Console.WriteLine("Row feature importance:");
// Order features by importance
var predictContributions = prediction.FeatureContributions
.Select((item, index) => new { index, item })
.OrderByDescending(theWeight => Math.Abs(theWeight.item));
foreach (var predictContribution in predictContributions)
{
Console.WriteLine($"Feature {featureColumns[predictContribution.index]} has feature contribution of {predictContribution.item}");
}
}
