Practical Applications to Timeseries Financial Data

Hi fellow Deep Learning enthusiasts,

I have finished this specialization and did all labs meticulously. However, I am not much interested in computer vision or NLP. I work in financial services sector and would like to be able to take the learning and apply them to financial timeseries data. I fail to use the code I have built using instructions of the labs, mostly in the fifth course, Sequence Models, and apply them to this particular domain. There are no issues with vanilla LSTM and GRU but when it comes to transformer networks and Encoder/Decoder, I struggle.

I know there are ample additional resources out there, and I have already tried and implemented some. However the labs are neat and some advice would be great on how to best think of modifying the codes written and applying those to timeseries financial data. Thank you very much all for your advice.

Additionally, if there are additional resources, particularly on application of advanced sequence models to timeseries data, I would be grateful to any recommendations.

1 Like

I am not good at Transformer, maybe @Juan_Olano could help you with this. But regarding LSTM, you may take the 4th course of TensorFlow Developer Professional Certificate named Sequences, Time Series and Prediction.