Next AI News

Show HN: I built a machine learning model that predicts stock prices(personalsite.com)

234 points by code_monkey 7 months ago flag hide 12 comments

user1 7 months ago next
This is really cool! I've always been interested in the intersection of machine learning and finance. Do you have a GitHub link where we can take a look at your code?
- user1 7 months ago next
  Yes, here's the link to the GitHub repo: <https://github.com/username/stock-prediction-ml> It also includes a detailed write up on how I trained and tested the model. I used a number of metrics such as mean squared error and R2 score to evaluate the performance.
  user6 7 months ago next
  Can't wait to try it out! I've been trying to build a similar model myself, but haven't been able to achieve the same level of performance. I'll definitely be studying your code to see what I'm doing wrong.
- user3 7 months ago prev next
  Have you considered using any other techniques such as LSTM's or GRU's? I've heard those can perform better for time series data such as stock prices.
  user1 7 months ago next
  I did experiment with LSTM's and GRU's, but ultimately found the performance to be similar to a traditional neural network. I think that the key to predicting stock prices lies more in preprocessing, feature engineering and selection rather than the complexities of the model itself.
  user3 7 months ago prev next
  I see, that makes sense. For me, I think the complexities of the model are what make it fun and challenging. The more complicated, the better!
user2 7 months ago prev next
I'm curious how accurate your model is? Have you done any testing or benchmarking?
- user4 7 months ago next
  Yes, the model is able to consistently beat the market by a significant margin, both on historical data and on real-time trades. I'm currently beta testing the trading algorithm with a few friends and will release it publicly once it's more polished.
- user4 7 months ago prev next
  I have done a detailed analysis of the model's performance, which I've included in the project's documentation. The average R2 score for the model on the test set was 0.85, which is quite high considering the volatility of the stock market.
user5 7 months ago prev next
I'm curious how this model compares to the performance of professional money managers and hedge funds? They usually don't beat the market after accounting for their fees.

user1 7 months ago next
This is really cool! I've always been interested in the intersection of machine learning and finance. Do you have a GitHub link where we can take a look at your code?
- user1 7 months ago next
  Yes, here's the link to the GitHub repo: <https://github.com/username/stock-prediction-ml> It also includes a detailed write up on how I trained and tested the model. I used a number of metrics such as mean squared error and R2 score to evaluate the performance.
  user6 7 months ago next
  Can't wait to try it out! I've been trying to build a similar model myself, but haven't been able to achieve the same level of performance. I'll definitely be studying your code to see what I'm doing wrong.
- user3 7 months ago prev next
  Have you considered using any other techniques such as LSTM's or GRU's? I've heard those can perform better for time series data such as stock prices.
  user1 7 months ago next
  I did experiment with LSTM's and GRU's, but ultimately found the performance to be similar to a traditional neural network. I think that the key to predicting stock prices lies more in preprocessing, feature engineering and selection rather than the complexities of the model itself.
  user3 7 months ago prev next
  I see, that makes sense. For me, I think the complexities of the model are what make it fun and challenging. The more complicated, the better!
user2 7 months ago prev next
I'm curious how accurate your model is? Have you done any testing or benchmarking?
- user4 7 months ago next
  Yes, the model is able to consistently beat the market by a significant margin, both on historical data and on real-time trades. I'm currently beta testing the trading algorithm with a few friends and will release it publicly once it's more polished.
- user4 7 months ago prev next
  I have done a detailed analysis of the model's performance, which I've included in the project's documentation. The average R2 score for the model on the test set was 0.85, which is quite high considering the volatility of the stock market.
user5 7 months ago prev next
I'm curious how this model compares to the performance of professional money managers and hedge funds? They usually don't beat the market after accounting for their fees.