A Novel Deep Convolutional Neural Network-based Trend Following Strategy for Stock Price Prediction

The stock market is complex and dynamic system and predicting the price of stock is a big challenge. Deep learning techniques provides deep convolutional neural network that has promising result in stock price prediction. This paper gives a novel deep convolutional neural network-based strategy for stock market trading and combining deep learning with the principle of trend following. It also addresses the limitation of previous models by analysing deep framework and incorporating reward function that has sustainable trading strategies. The model outperforms baseline method on key performance indicators such as annualized return and risk adjusted returns demonstrating the potential of deep learning in stock market trading.