Stock Price Prediction with CNN-LSTM Network Yushi Guan Peiyao Li Cheng Lu Abstract Stock price prediction has always been a rewarding but challenging task since are inherent noises present in the stock time series data. The application of machine learning and deep learning in this domain has attracted many computer science researchers’ attention. Most published state of the art machine learning models for stock prediction use a large set of technical and economic factors as feature inputs along with the target stock’s historical data to predict future price movements [2][3][9]. Inspired by recent advances in natural language processing and speech synthesis [7][22], we propose a combined CNN-LSTM model that can achieve state of the art performance for stock price prediction without additional data such as technical and economic indicators in the input features. The proposed model outperformed a simpler LSTM model as well as non-neural methods implemented such as RF, KNN, and linear regression. Using a simple trading strategy, we demonstrate that our stock price prediction algorithm outperforms the buy and hold strategy on various index and stocks including SP500 index ETF (SPY), Intel (INTC), Tesla (TSLA). The SPY, INTC, TSLA, and M achieved Sharpe ratios of 0.74, 1.73, and 1.22 respectively. https://erikloo.github.io/files/Stock Price Prediction with CNN-LSTM Network.pdf