keras style wrapper for N_Beats time series prediction model library, plot prediction data, get prediction data, improved errors and guardrails for simple prediction use.
Download directions are at the bottom of the page
import ezNbeats as Nbeats
Nbeats.run_prediction_and_plot('ethdata.csv', 'period_prediction_in_integer_value')
Returns:
Nbeats.return_prediction('ethdata.csv', 'period_prediction_in_integer_value')
Returns:
array([[4533.275 ],
[4763.67 ],
[3268.3918],
[4135.316 ],
[4364.2817],
[2880.5466],
[4336.967 ],
[4308.3403],
[4175.3003],
[4983.2505],
[4494.187 ],
[4606.2715],
[4599.9175],
[4616.081 ],
[5306.367 ],
[5355.2334],
[5408.6665],
[5590.524 ],
[4804.6304],
[5413.8545],
[5110.1167],
[4739.2515],
[5725.405 ],
[4756.4424],
[5187.714 ],
[5316.915 ],
[5154.975 ],
[4508.4106],
[4495.5225],
[4979.912 ],
[6297.265 ],
[4575.6895],
[4595.218 ],
[4784.8228],
[4837.4814],
[4257.399 ],
[4152.7993],
[5176.271 ],
[5433.6543],
[6237.472 ],
[5223.8813],
[5309.657 ],
[5729.7686],
[5129.54 ],
[6764.619 ],
[7174.92 ],
[6157.9614],
[6530.81 ],
[6562.5435],
[6785.9463],
[6472.0635],
[6528.72 ],
[5347.449 ],
[6292.9966],
[5805.043 ],
[6684.3423],
[6143.1997],
[5369.238 ],
[6882.697 ],
[5556.365 ],
[5842.768 ],
[5629.788 ],
[6172.0137],
[5913.5005],
[5939.173 ],
[5678.9614],
[5247.0776],
[5601.757 ],
[6229.9883],
[6342.4746]], dtype=float32)
Clone repo
Cd into repo in terminal
pip3 install .
in python file:
import ezNbeats as Nbeats
dependencies are tough as pytorch, sci libraries, and nbeats_forecast all require different versions of numpy and etc. Recommended install in a venv, and add dependecies one by one, finally install nbeats_forecast as pip3 install nbeats_forecast --no-cache-dir
Contact me if you have issues downloading