Kapurvish you may use multiple stock prediction using single neural network. I have experimented with multiple stock e.g. list = [ “IBM”, “MSFT”, “GOOGL”, “GE”, “AAPL”, “AMZN”, “CSCO”]. However, from my experience, even though neural networks are quite powerful, they suffer from vanishing gradient issues. You may read through this. In LSTM the short-term memory can last for a long period of time. LSTMs can deal with the exploding and vanishing gradient problem when training traditional RNNs.