/trendPy

Time Series Regression with Python

Primary LanguageJupyter NotebookMIT LicenseMIT

DOI Jupyter Lab: Binder Documentation: doc WebApps: Binder (binder) Streamlit App

Usage

pip install trendpy2

and use it as trendpy2 as shown in the example.ipynb and approximate your time series ($f:\mathbb{R}\to \mathbb{R}$) with the following trends

  • linear $f(x)=a\cdot x+b$
  • polynomial $f(x)=a_n\cdot x^n+a_{n-1}\cdot x^{n-1}+...+a_0$
  • exponential $f(x)=a\cdot e^{b\cdot x}$
  • trigonometric $f(x)=a\cdot \cos(2\cdot \pi\cdot b\cdot x+c)$
  • "free" (for max. three parameters) (e.g.a*arctan(b*x+c), a*exp(b*x+c), a*(x*b)+c), the intial guess for a, b, c is 1.

in your Python scripts or jupyter notebooks and use the best of the numerical and symbolic worlds to make predictions and assess the quality of your fit!

trendpy2 is deterministic, i.e. complementary to trendpy, which uses a stochastic approach.

or use one of the WebApps with the correspondig button above (voila app or streamlit app).

For more, have a look at the sphinx-documentation!

Voila App

Streamlit App