Predicting Bitcoin Price using a Recurrent Neural Network (RNN)

I have been experimenting and learning about Recurrent Neural Networks (RNNs).  In the course that I took, I trained a model to predict Google's stock price.  Based on the success of the model, I decided to use the same program to try to predict Bitcoin's price from January 1, 2023 - February 28, 2023.


Here is the info:

Recurrent Neural Network (LSTM)

Training on 60 days of past price

Predicting 1 day out

Training data: June 20, 2021 - Dec 31, 2022

Training epochs : 100

Training mean square error: loss: 0.0024

Testing data: January 1, 2023 - February 28, 2023

Data source: Coingecko.com


Here is the chart detailing the actual Bitcoin price with the RNN predictions.

(Note: The chart is in French as I am taking a deep learning course in French. I will be producing more charts so future charts will be in English)






Comments

Post a Comment

Popular posts from this blog

Coingecko Error: Exceded Rate Limit

Image
Bad news again today! I received an error message from CoinGecko for exceeding the rate limit.  I have reduced my daily request from 10,000 cryptos to 5,000 cryptos, and -after some time- I still received an error message.   I have now reduced my request to 2,500 cryptos.  I believe that my error messages are a result of downloading a large number of cryptos in a 24-hour period. Hopefully, my smaller sample size will overcome this problem.
Read more

Reducing my request from Conigecko

I have now reduced my request to 1,000 cryptos per day.  I believe that my error messages are a result of downloading a large number of cryptos in a 24-hour period over a number of days. Hopefully, my smaller sample size will overcome this problem.
Read more

Is Solana dead? Thoughts from ChatGPT.

  Many people think that Solana is dead because of the numerous network outages.  Is Solana dead?  Let's ask ChatCPT.... Q : Why does Solana still keep having problems with its network? ChatGPT: While Solana is designed to be a high-performance and scalable blockchain platform, it has experienced several network issues in the past. Some of the reasons for these problems include: 1) Network congestion: As more applications and users have joined the Solana network, the network has become congested, leading to delays and slower transaction processing times. This has resulted in higher fees and increased the risk of network congestion-related issues. 2) Centralization of network validators: The Solana network currently relies on a limited number of validators, which are responsible for validating transactions and maintaining the integrity of the blockchain. This centralization has raised concerns about the security and decentralization of the network, particularly if the...
Read more