Timeseries BitPredict Exercise

[SadanandaJaya]

Is there a better way to create this model? As in, should the block_reward feature be bundled in with the Bitcoin historical price feature? Perhaps you could test whether building a multi-input model (e.g. one model input for Bitcoin price history and one model input for block_reward) works better? See Model 4: Hybrid embedding section of notebook 09 for an idea on how to create a multi-input model.

for the above exercise, i have to make train windows and labels on the prices but for "block_reward" on which column should i make the train,test split.Has anyone solved this problem is yes, please do share the code it will be a great help.

[mustafa2mujahid]

Have you found the answer to this?