First let me say, I wouldn’t be as computer literate as some, but I try.
“Anyways and all that” as the saying goes.

Quick short history, about the oldest coin in the world.
The heritage of ancient coins is a subject that intrigues and delights collectors and scholars the world over. The oldest coin available today was discovered in Efesos, an ancient Hellenic city and prosperous trading center on the coast of Asia Minor. The 1/6 stater, pictured below, is more than 2,700 years old, making it one of the very earliest coins. Made from electrum, a natural occurring alloy of gold and silver, the coin originated in the area of Lydia. It had a design on one side only, a result of the primitive method of manufacture. This ancient stater was hand struck. A die with a design (in this case a lion's head) for the obverse (front) of the coin was placed on an anvil. A blank piece of metal was placed on top of the die, and a punch hammered onto the reverse. The result was a coin with an image on one side and a punch mark on the other.

ref: https://www.fleur-de-coin.com/articles/oldest-coin

As I said It’s only natural someone would do this eventually, so here’s my contribution for what it’s worth, a token if you will, it features the Lydian kings' emblem of a roaring lion, almost always with a curious knob, often called a "nose wart," on its forehead.

Hey Siraj, thanks for the tutorial!

I can't find a license attached to this repo - is the code open-source/available for re-use in other projects?

Hi Siraj,
I've experimented with RNN/NN/CNN/.. and crypto price predictions in the past and most of the time these models just learned the global statistical distribution of the dataset, because in the short term the random walk model only applies and nothing more.
To illustrate this, I build a simple python script which just samples the Data according to a Gaussian distribution and pulls some samples from it to evaluate this simple price prediction Model.
The result is that this simple "Model" has a similar (in the interval of mean+2*std) Precision,Recall and F1 Score as the RNN with just the Price as Input and nothing else.
Best regards,
Julien

Result Simple Gaussian Model:
Mean Precision: 0.5823 | Mean Recall: 0.5561 | Mean F1 Score: 0.5686
STD Precision: 0.0227 | STD Recall: 0.0348 | STD F1 Score: 0.0266

Where could I get the data for this model? The data in the Kaggle competition referred to in the video is locked. The hash and block metrics for the coins are available but some of the features look like they could be engineered. Is this correct?

No this csv, it is hard to say these code workable... Pls upload.

Can someone explain why the data is converted into a 3D array and also why we shuffle? Wouldn't this data be a time series so the order would matter?

#Convert the data to a 3D array (a x b x c)
#Shuffle the data np.random.shuffle(training_data)

I assume based on the outline the data would have an index for the day and then have a value for each of the fields he outlined but not sure why the time sequence wouldn't take a part in this and we would shuffle? My thought on the data format to feed into this would be a CSV with each feature value and the target being price but little confused.

Hey!

Is it better to normalize over a single window or over the full data set? Should the test set included? Could you tell me why?

Thanks,
Philmod

Did you ever think about running this? Or was this for the lols

"I love your videos Siraj, but this is very disappointing. Using a bidirectional network is beyond cheating. If you were to actually try to use this, the network requires past data AND future data as inputs to make a prediction, which won't exist yet. Also even if you use a regular LSTM, this type of scheme doesn't work. The network learns to predict the input at time t plus some noise as the output. It creates nice looking graphs but is useless in trading. I know this is a topic that will get you some views but the least you can do is mention that it is cheating and won't work at all before you waste peoples' time."

In the referenced video, we are told that you can find a dataset on Kaggle.
The datasets which can be found there does not contain all the columns that this notebook requires.

So where can we find a dataset that can be used together with this model?

Traceback (most recent call last): File "C:\Users\Admin\Desktop\btcpy.py", line 328, in X_train, Y_train, X_test, Y_test, Y_daybefore, unnormalized_bases, window_size = load_data("bitcoinprices.csv", 50) File "C:\Users\Admin\Desktop\btcpy.py", line 67, in load_data unnormalized_bases = d0[start:end,0:1,21] IndexError: index 21 is out of bounds for axis 2 with size 2

This error i got when i run the program nd is unable to solve the error so help me with this