This paper is about predicting the movement of stock consist of S&P 500 index. Historically there are many approaches have been tried using various methods to predict the stock movement and being used in the market currently for algorithm trading and alpha generating systems using traditional mathematical approaches [1, 2].
The success of artificial neural network recently created a lot of interest and paved the way to enable prediction using cutting-edge research in the machine learning and deep learning. Some of these papers have done a great job in implementing and explaining benefits of these new technologies. Although most these papers do not go into the complexity of the financial data and mostly utilize single dimension data, still most of these papers were successful in creating the ground for future research in this comparatively new phenomenon. In this paper, I am trying to use multivariate raw data including stock split/dividend events (as-is) present in real-world market data instead of engineered financial data. Convolution Neural Network (CNN), the best-known tool so far for image classification, is used on the multi-dimensional stock numbers taken from the market mimicking them as a vector of historical data matrices (read images) and the model achieves promising results. The predictions can be made stock by stock, i.e., a single stock, sector-wise or for the portfolio of stocks.
The Standard & Poor's 500 Index or S&P 500 as it called is a weighted index of the 500 largest U.S. publicly traded companies market capitalization. S&P 500 is one of the furthermost commonly quoted American indexes because it is representative of the largest U.S. public corporations and it focuses on the large-cap sector of the U.S. market. [3] Financial institutions, banks, dealer broker firms, hedge funds and individual rely on S&P 500 stock's movement to invest, liquidate their investments or hedge their business risks. S&P 500 represent top 500 large cap companies of the United States and is watched by companies and individual not just in the US but around the globe as one of the key financial indicators for market movements. Though the same proposed model can be trained for all the stocks listed in S&P 500, for illustration purposes, in this paper the proposed model will use historical prices of few of S&P 500 stocks to train and test the model.
In the recent years, deep learning using Artificial Neural Networks (ANN) have been used in a lot of different areas such, to name few such as automotive, diagnostic, surgery, etc. ANN application list is ever increasing. After the ImageNet challenges where Convolutional Neural Network (CNN) based model have made significant progress in predicting images (in fact better than human) and overdo other available approaches used till date, there have been several applications tried using CNN. Another Deep Learning approach, LSTM (stands for Long-Short-Term-Memory), a prominent variation of Recurrent Neural Network or RNN, has been used in predicting time-series data and has provided surprising results. A recent study shows that LSTM outperforms traditional-based algorithms such as ARIMA model [4] .
The application of deep learning approaches mainly LSTM and CNN to finance has received a great deal of attention from both investors and academia. LSTMs are challenging to train, mainly because of the computationally expensive nature of time-based gradient descent, the size of the networks, and the amount of data over which they must be trained. Further complicating, since LSTMs are stateful, many problems require “online” training, meaning that they cannot be trained all at once by highly optimized, vectorized calculations in a batch with error computation and gradient descent over groups [5] . However, CNNs with
Rahul Gupta Stanford University 353 Serra Mall, Stanford, CA 94305 rahulgup@stanford.edu their ability to learn useful spatial features from the input data have revolutionized computer vision. The network topology of CNNs exploits the spatial relationship among the pixels in an image, and this is one of the reasons for their success [6] .
Stock market prediction is usually considered as one of the most challenging issues among time series predictions [7] due to its noise, and volatile features make CNN natural choice over LSTM. In addition to the data noise, due to its nature of convolving CNN intuitionally would be better of handling prediction of a stock dividend and split events.
The input to our algorithm is a stock’s raw historical time-series numbers (OPEN, HIGH, LOW, CLOSE, VOLUME, ADJ_OPEN, ADJ_HIGH, ADJ_LOW, ADJ_CLOSE, ADJ_VOLUME) uploaded in CSV format for all current stock consists in S&P 500. We then use a CNN approach to deep learning to predict BUY/SELL based on the absolute return.
Persio and Honchar [8] laid the framework for comparing different neural network models for stock prediction, concerning the forecast of their trend movements up or down, in their paper Artificial Neural Networks architectures for stock price prediction: comparisons and applications. Like our data set, the S&P 500 historical time series, predicting a trend based on data from the past days, and proposing a novel state-of-art approach based on a combination of wavelets and CNN, which outperforms the basic neural networks ones. Their paper demonstrates that neural networks can predict financial time series movements even trained only on plain time series data.
In their paper, Financial Time Series Prediction using Deep Learning, Ariel Navon, and Yosi Kellery [9] presented a data-driven end-to-end Deep Learning approach for time series prediction, applied to financial time series. In their paper, a Deep Learning scheme is derived to predict the temporal trends of stocks and ETFs in NYSE or NASDAQ. Their approach is based on an artificial neural network to predict the temporal patterns of stocks and ETFs that are applied to raw financial data inputs which very similar to the input data being used in this paper as it used raw market data of indices/ticker from S&P 500.
Convolutional neural networks (CNNs) with their ability to learn useful spatial features have revolutionized computer vision. In Predicting Time Series with Space-Time Convolutional and Recurrent Neural Networks, Grob et al. applied CNNs to predict electricity prices at the European Power Exchange dis
This content is AI-processed based on open access ArXiv data.
