Deep learning for sequence pattern recognition

[DrR0bot]

Hello everyone!

I am highly focused on learning the application of RNN for sequence pattern recognition. I am starting week 3 of DLS course five, and this kind of application is not within the course content.

I wanted to ask if anyone could supply me with educational resources, tutorials, examples or anything else that can help me build up a model for pattern recognition.

Thanks all for your help

[Christian_Simonis]

Hello,

in case you are interested in pattern modelling for Time Series analysis:

RNNs

• CS 230 - Recurrent Neural Networks Cheatsheet
• http://cs231n.stanford.edu/slides/2017/cs231n_2017_lecture10.pdf

Personally I find the tensorflow documentation very helpful, too:

• Recurrent Neural Networks (RNN) กับ Keras  |  TensorFlow Core

Good benchmarks:
ARIMA models might be worth a look since you can describe seasonal patterns and trends:

• Time series Forecasting — ARIMA models | by Sangarshanan | Towards Data Science
• Model selection with cross-validation — pmdarima 0.0.0 documentation

Another nice benchmark for an automatic time series modelling can be derived with TSFresh:

• tsfresh — tsfresh 0.18.1.dev39+g611e04f documentation

Please let me know if this goes in a helpful direction for you.