96 points by ms_trend_predictor 6 months ago flag hide 9 comments
deeplearningmaster 6 months ago next
Fascinating article! I've been exploring similar approaches in my research lately. Has anyone played around with LSTM networks for predicting stock market trends?
datavizguru 6 months ago next
Yes, I've experimented with LSTM networks for stock market prediction! I found that preprocessing the data with a performant feature engineering method significantly improved the model's performance. @DeepLearningMaster
deeplearningmaster 6 months ago next
I've used feature scaling techniques such as min-max scaling for normalizing data while keeping it stationary. @DataVizGuru. I've not tried EDA yet, but will definitely check that out. Thanks, @StatsGuru!
quantml 6 months ago prev next
I think feature engineering is key when applying ML algorithms to stock market data. Any recommendations for that? @DeepLearningMaster @DataVizGuru
featureeng 6 months ago next
Definitely recommend checking out more advanced feature engineering methods like binning, polynomial interaction features, and hashing. @QuantML
featureeng 6 months ago next
Preventing overfitting is crucial. You can try using regularization, dropout, or further segmenting the data for cross-validation @TradingBot
statsguru 6 months ago prev next
Just came across Exploratory Data Analysis (EDA) techniques in ML for stock market prediction. It's worth a look! @DeepLearningMaster
tradingbot 6 months ago prev next
I've built a trading bot based on ML stock price prediction models. It's made me money, but I've also noticed significant losses because of overfitting. How can we prevent this in ML models? @DeepLearningMaster
rxforecaster 6 months ago prev next
Also interested in this thread. Would love to learn more about recurrent neural networks (RNN) and how they apply to stock price prediction. @DeepLearningMaster