Comparing the effectiveness of deep feedforward neural networks and shallow architectures for predicting stock price indices

Abstract

Many existing learning algorithms suffer from limited architectural depth and the locality of estimators, making it difficult to generalize from the test set and providing inefficient and biased estimators. Deep architectures have been shown to appropriately learn correlation structures in time series data. This paper compares the effectiveness of a deep feedforward Neural Network (DNN) and shallow architectures (e.g., Support Vector Machine (SVM) and one-layer NN) when predicting a broad crosssection of stock price indices in both developed and emerging markets. An extensive evaluation is undertaken, using daily, hourly, minute and tick level data related to thirty-four financial indices from 32 countries across six years. Our evaluation results show a considerable advantage from training deep (cf. shallow) architectures, using a rectifier linear (RELU) activation function, across all thirty-four markets when 'minute' data is used. However, the predictive performance of DNN was not significantly better than that of shallower architectures when using tick level data. This result suggests that when training a DNN algorithm, the predictive accuracy peaks, regardless of training size. We also examine which activation function works best for stock price index data. Our results demonstrate that the RELU activation function performs better than TANH across all markets and time horizons when using DNN to predict stock price indices.

An AI‐based model for describing cyclic characteristics of granular materials

“…Tanh and ReLU are two commonly used activation functions. Because the combination of two functions can overcome their own deficiencies, they are thus adopted in this study with detailed introduction in Appendix A (Figure S1).…”

Prediction-Based Portfolio Optimization Models Using Deep Neural Networks

“…DMLP contains many hyperparameters, this paper uses grid research method to discover its optimal hyperparameters. Since relu function performs better than tanh function [34], this paper applies relu function as activation function. All the considered hyperparameters are presented in Table 1.…”

“…And, earlystopping technology is used to reduce overfitting problem. Since relu function outperforms tanh function [34], relu function is adopted as activation function. After many trial and error, hidden node is set to 5, hidden layer is set to 1, learning rate is set to 0.001, patient is set to 0, batch size is set to 100, dropout rate is set to 0.1, recurrent dropout rate is set to 0.2, optimizer is set to RMSProp.…”

“…SVR is a classic machine learning technique, which has been widely used in stock price prediction [15], [34]. SVR is a nonlinear kernel-based regression method which tries to locate a regression hyperplane with small risk in highdimensional feature space.…”

Using machine learning techniques to predict the cost of repairing hard failures in underground fiber optics networks

“…Feedforward neural networks fundamentally consist of three layers, which are the input layer, hidden layer, and output layer. The FFNN model has one hidden layer with monotonically increasing differentiable functions, which has the ability to approximate the continuous function with the hidden layer [12]. The FFNN model used in this paper has two input neurons and a hidden layer with two nodes, which comprise of the data collected from the field, which are the determinants of the cost of fiber cable repairs.…”