Regression

Learn how to predict continuous values using AiDotNet’s regression algorithms.

What is Regression?

Regression is a supervised learning task where the goal is to predict continuous numeric values from input features. Examples include:

• House price prediction
• Temperature forecasting
• Stock price estimation
• Sales revenue prediction

Types of Regression

Simple Regression

One input feature predicts one output. Example: Predict house price from square footage.

Multiple Regression

Multiple input features predict one output. Example: Predict house price from square footage, bedrooms, and location.

Polynomial Regression

Non-linear relationships using polynomial features.

Quick Start

using AiDotNet;
using AiDotNet.Regression;

// Prepare data (house features: sqft, bedrooms, bathrooms, age)
var features = new double[][]
{
    new[] { 1400.0, 3.0, 2.0, 15.0 },
    new[] { 1600.0, 3.0, 2.0, 10.0 },
    new[] { 1700.0, 3.0, 2.5, 5.0 },
    new[] { 1875.0, 4.0, 3.0, 8.0 },
    new[] { 1100.0, 2.0, 1.0, 25.0 },
    new[] { 2200.0, 4.0, 3.0, 2.0 }
};
var prices = new double[] { 245000, 312000, 279000, 308000, 199000, 425000 };

// Build and train
var result = await new AiModelBuilder<double, double[], double>()
    .ConfigureModel(new RandomForestRegressor<double>(nEstimators: 100))
    .ConfigurePreprocessing()
    .BuildAsync(features, prices);

// Predict
var newHome = new double[] { 1500.0, 3.0, 2.0, 12.0 };
var predictedPrice = result.Predict(newHome);
Console.WriteLine($"Predicted price: ${predictedPrice:N0}");

Available Regressors

Tree-Based Methods

Linear Methods

Distance-Based

Neural Networks

var result = await new AiModelBuilder<float, Tensor<float>, Tensor<float>>()
    .ConfigureModel(new NeuralNetworkRegressor<float>(
        inputFeatures: 10,
        complexity: NetworkComplexity.Medium))
    .ConfigureOptimizer(new AdamOptimizer<float>(learningRate: 0.001f))
    .BuildAsync(features, targets);

Data Preprocessing

Automatic Preprocessing

.ConfigurePreprocessing()  // Applies StandardScaler by default

Custom Preprocessing

.ConfigurePreprocessing(new PreprocessingConfig
{
    Scaler = new MinMaxScaler<double>(),
    ImputeStrategy = ImputeStrategy.Median,
    HandleCategorical = true
})

Evaluation Metrics

var predictions = testSamples.Select(s => result.Predict(s)).ToArray();

// Common regression metrics
Console.WriteLine($"MAE: {Metrics.MeanAbsoluteError(predictions, testTargets):F4}");
Console.WriteLine($"MSE: {Metrics.MeanSquaredError(predictions, testTargets):F4}");
Console.WriteLine($"RMSE: {Metrics.RootMeanSquaredError(predictions, testTargets):F4}");
Console.WriteLine($"R2: {Metrics.RSquared(predictions, testTargets):F4}");

Best Practices

1. Start simple: Use Linear Regression as a baseline
2. Check for outliers: Outliers heavily influence linear models
3. Feature scaling: Most algorithms benefit from scaled features
4. Regularize: Use Ridge or Lasso to prevent overfitting
5. Validate: Use cross-validation for robust evaluation

Next Steps

• Classification Tutorial - For predicting discrete labels
• Time Series Tutorial - For temporal data