Time series prediction

[Daniel Wee]

Some common algorithms - Rain Forest, SVR (Support Vector Regression), ANN

https://github.com/EdjeElectronics/Tens ... Windows-10
https://www.youtube.com/watch?v=fNmJJq2Enqo

[Daniel Wee]

https://www.youtube.com/watch?v=IscG_bAeEic
http://amunategui.github.io/

[Daniel Wee]

Multivariate time-series, single and multi-step implementation:-

https://www.tensorflow.org/tutorials/st ... ime_series

[Daniel Wee]

Some useful descriptions in here:-

https://medium.com/google-cloud/how-to- ... d2eeb189e8

[Daniel Wee]

Looks like the Transformer architecture is better than the LSTM which is better, in turn, than RNNs, for handling time dependencies.

https://medium.com/tensorflow/a-transfo ... bf59e66fe2

[Daniel Wee]

A good article on the various considerations on the prediction power of a neural model.

https://www.kdnuggets.com/2019/05/machi ... sting.html
https://www.kdnuggets.com/2020/03/machi ... equel.html

Multi-variate time-series forecasting:-

https://www.analyticsvidhya.com/blog/20 ... hon-codes/
https://machinelearningmastery.com/how- ... recasting/
https://machinelearningmastery.com/mult ... tms-keras/

https://towardsdatascience.com/var-and- ... b8d8888141
https://towardsdatascience.com/multivar ... 3372b3db36

https://www.statsmodels.org/dev/vector_ar.html
https://jacksonpradolima.github.io/PyNe ... al_nn.html

https://pages.databricks.com/rs/094-YMS ... Usage.html
https://www.mariofilho.com/how-to-predi ... g-example/

[Daniel Wee]

Interesting github projects:-

https://github.com/maxim5/time-series-machine-learning
https://github.com/microsoft/forecasting

Cryptocurrency datasource:-

https://bitcoincharts.com/charts/

[Daniel Wee]

The sktime library also offers some useful tools that include auto-regressive predictions (multi-step recursive).

[Daniel Wee]

https://github.com/microsoft/forecastin ... ound.ipynb