Author: Alex Spanos
Date: 09/08/18
- Latest version:
0.1.8dev
This repository contains helper code for time series forecasting problems.
The aim for the package is to provide a unified interface for tackling time series forecasting problems, by allowing a quick and easy performance comparison for standard methodologies such as the naive method, exponential smoothing, ARIMA and standard machine learning.
timeseries_toolkit requires
- Python (3.6)
numpypandasscipy
pip install -U timeseries-toolkit
Note: if installation fails on MacOSX because of pip failing to
build wheels for psutil, the reason is most likely due to a gcc
incompatibility. Doing conda install gcc before trying to install
pip install -U timeseries-toolkit should fix the problem.