Abstract:Stock market prediction is a very hot topic in financial world. Successful prediction of stock market movement may promise high profits. However, an accurate prediction of stock movement is a highly complicated and very difficult task because there are many factors that may affect the stock price such as global economy, politics, investor expectation and others. Several non-linear models such as Artificial Neural Network, fuzzy systems and hybrid models are being used for forecasting stock market. These models have limitations like slow convergence and overfitting problem. To solve the aforementioned issues, this paper intends to develop a robust stock trading model using deep learning network. In this paper, a stock trading model by integrating Technical Indicators and Convolutional Neural Network (TI-CNN) is developed and implemented. The stock data investigated in this work were collected from publicly available sources. Ten technical indicators are extracted from the historical data and taken as feature vectors. Subsequently, feature vectors are converted into an image using Gramian Angular Field and fed as an input to the CNN. Closing price of stock data are manually labelled as sell, buy, and hold points by determining the top and bottom points in a sliding window. The duration considered over a period from January 2009 to December 2018. Prediction ability of the developed TI-CNN model is tested on NASDAQ and NYSE data. Performance indicators such as accuracy and F1 score are calculated and compared to prove effectiveness of the proposed stock trading model. Experimental results demonstrate that the proposed TI-CNN achieves high prediction accuracy than that of the earlier models considered for comparison.

Predict Forex Trend via Convolutional Neural Networks

Convolutional neural network for stock trading using technical indicators

Applying Convolutional Neural Networks for Stock Market Trends Identification

Algorithmic financial trading with deep convolutional neural networks: Time series to image conversion approach

Stock Market Trend Prediction Using Recurrent Convolutional Neural Networks.

Deep learning-based predictive models for forex market trends: Practical implementation and performance evaluation

Forecasting Foreign Exchange Market Prices Using Technical Indicators with Deep Learning and Attention Mechanism

Financial Markets Prediction with Deep Learning

A novel graph convolutional feature based convolutional neural network for stock trend prediction

Deep Learning for Price Movement Prediction Using Convolutional Neural Network and Long Short-Term Memory

Research on a Trend Forecast Model of Time Series Data Based on Deep Learning

Financial Trading Model with Stock Bar Chart Image Time Series with Deep Convolutional Neural Networks

A Novel Convolutional Neural Networks for Stock Trading Based on DDQN Algorithm

Graphical Deep Learning Prediction Model for Stock Risk Management

Stock Market Trend Prediction Using High-Order Information of Time Series

MG-Conv: A spatiotemporal multi-graph convolutional neural network for stock market index trend prediction

Do Deep Learning Models and News Headlines Outperform Conventional Prediction Techniques on Forex Data?

Classification-based Financial Markets Prediction using Deep Neural Networks

Thresholded ConvNet Ensembles: Neural Networks for Technical Forecasting

Important Trading Point Prediction Using a Hybrid Convolutional Recurrent Neural Network

Deep Convolutional Transformer Network for Stock Movement Prediction