nemreg 0.1.0

An easy-to-use Python regression library.

nemreg

A lightweight regression framework built on top of SciPy.

nemreg is designed to make regression analysis structured, compact, and easy to use — without sacrificing flexibility.
It wraps scipy.optimize.curve_fit in a clean scientific workflow that includes dataset handling, statistics, summaries, and plotting.

---

Install nemreg

install nemreg

Why nemreg?

SciPy is extremely powerful — but low-level.
With SciPy alone, you must manually:

• Prepare data shapes
• Compute residuals
• Compute R² / RMSE / χ²
• Handle uncertainties
• Write plotting code
• Structure outputs

nemreg provides this structure for you.

import nemreg as nr

ds = nr.Dataset(x, y, yerr=1.0)
res = nr.fit(ds, nr.models.linear())
nr.plot(mode="result_data", dataset=ds, result=res)

That’s it.

---

Features

• Clean Dataset abstraction (1D and multivariable)
• Weighted and unweighted regression
• Automatic statistics:
  • R²
  • RMSE
  • χ²
  • Reduced χ²
• Structured FitResult object
• Built-in plotting dispatcher
• Built-in models (linear, polynomial, etc.)
• Support for replicated measurements
• Clean summary output
• Multivariable regression support

---

Installation

Clone the repository and install dependencies:

pip install numpy scipy matplotlib

Optional:

pip install pytest jupyter

Then install nemreg (editable mode recommended):

pip install -e .

---

Quick Start

import numpy as np
import nemreg as nr

x = np.linspace(-5, 5, 40)
y = 2.5*x - 1 + np.random.normal(0, 1, size=x.size)

ds = nr.Dataset(x=x, y=y)
res = nr.fit(ds, nr.models.linear())
nr.plot(mode="result_data", dataset=ds, result=res)

---

Project Philosophy

nemreg is not meant to replace SciPy.

It is meant to:

• Reduce boilerplate
• Encourage structured analysis
• Improve readability
• Make regression workflows easier to teach and share

SciPy remains the optimization engine under the hood.

---

Note from the Author

This library is intended for free use and free sharing, as it was created in the spirit of making regression easier for everyone.

Please refer to the LICENCE file for details on usage, modification, and redistribution.

The author strongly recommends browsing through the nemreg_demo.ipynb file.
It allows you to:

• Verify dependencies
• Confirm your environment works correctly
• See example workflows
• Explore available commands

---

Planned Features

• LaTeX export for automatic report generation
• Native Levenberg–Marquardt (LM) core implementation, trying to improve the regression capabilities on more complex functions
• Additional diagnostics (influence, leverage, etc.)
• Robust fitting / outlier handling
• Extended model library
• Improved multivariable visualization

---

License

See the LICENCE file for full details.

---

Contributing

Contributions, suggestions, and improvements are welcome.
Please open an issue or submit a pull request.

---

Acknowledgements

Built on top of:

• NumPy
• SciPy
• Matplotlib

These libraries power the numerical engine behind nemreg.