Datasets to incl:

• [x] m4 daily
• [ ] gunpoint
• [ ] Phoneme http://www.timeseriesclassification.com/description.php?Dataset=Phoneme
• [ ] Worms http://www.timeseriesclassification.com/description.php?Dataset=Worms
• [ ] BeetleFly http://www.timeseriesclassification.com/description.php?Dataset=BeetleFly
• [ ] m4 hourly

using zip files https://forums.fast.ai/t/generalizing-untar-data-to-also-work-with-zips/53741/14?u=takotab

URLs.m4_daily

'https://motionnet-m4-dataset.s3.eu-central-1.amazonaws.com/m4_daily.tgz'

dummy_data_generator[source]

dummy_data_generator(lookback:int, horizon:int, signal_type='seasonality', nrows:int=5, random=True, batch_size=32, norm=False, noise=0.2)

r = dummy_data_generator(30,10,signal_type='trend' )
plt.plot(r[0][0,:])
plt.plot(r[1][0,:])
plt.plot(r[2][0,:])
plt.plot(r[3][0,:])

[<matplotlib.lines.Line2D at 0x7f3d82156090>]

r = dummy_data_generator(30,10)
plt.plot(r[0][0,:])
plt.plot(r[1][0,:])
plt.plot(r[2][0,:])
plt.plot(r[3][0,:])

[<matplotlib.lines.Line2D at 0x7f3d81f4f1d0>]

r = dummy_data_generator(30,10,norm=True, noise = 0)
plt.plot(r[0][0,:])
plt.plot(r[1][0,:])
plt.plot(r[2][0,:])
plt.plot(r[3][0,:])

[<matplotlib.lines.Line2D at 0x7f3d81bb8590>]