make_regression_check_data: Synthetic Regression Dataset

The make_regression_check_data function generates synthetic regression datasets. It supports linear, polynomial, and sinusoidal relationships with additive Gaussian noise and optional outliers. This serves as a versatile testbed for validating regression models within the Machine Gnostics framework.

---

Overview

This utility creates regression datasets for testing:

• Functions: Linear/Polynomial (\(y = \sum a_i x^i + c\)), Sinusoidal (\(y = A \sin(x) + c\)), or Cosine.
• Outliers: Option to contaminate a percentage of data with large shifts (outlier_ratio), ideal for testing robust regression capabilities.
• Purpose: Unit testing models, verifying pipeline integrity, and demonstrating regression capabilities.
• Reproducibility: Uses a fixed seed (default 42).

---

Parameters

Parameter	Type	Description	Default
n_samples	int	The number of data points to generate.	20
slope	float	True coefficient (linear/poly) or amplitude (sin/cos).	3.5
intercept	float	The true intercept (bias) of the underlying relationship.	10.0
noise_level	float	Standard deviation of Gaussian noise added to target.	2.0
degree	int	Degree of the polynomial. Used if function_type='poly'.	1
function_type	str	Component function type: 'poly', 'sin', 'cos'.	'poly'
outlier_ratio	float	Proportion of samples to contaminate with outliers (0.0 to 1.0).	0.0
seed	int	Random seed for reproducibility.	42

---

Returns

Return	Type	Description
X	numpy.ndarray	Input feature array of shape (n_samples,). Values uniform in [0, 10].
y	numpy.ndarray	Target array of shape (n_samples,).

---

Example Usage

from machinegnostics.datasets import make_regression_check_data
import numpy as np

# Example 1: Standard Linear Regression
X, y = make_regression_check_data(n_samples=50)

# Example 2: Sinusoidal data with outliers
X_sin, y_sin = make_regression_check_data(
    n_samples=100, 
    function_type='sin', 
    outlier_ratio=0.1
)

print(f"X shape: {X.shape}")
print(f"y shape: {y.shape}")

---

Author: Nirmal Parmar