Systematic Equity Alphas Library
Find out our latest insights in systematic equity alphas

Financial News to Predict Stock Market
Financial news provides information to the general public. Consumers rely on information they read or hear before buying a product.
Read More


Support Vector Machine for Stock Market Prediction
Predicting direction of stock price index movement using artificial neural networks and support vector machines
Read More


Target price forecasts Fundamentals and behavioral factors
Group Normalization provides a more stable data normalization methodology that can ease optimization and enable very deep networks to converge.
Read More


Accrual Reliability, Earnings Persistence and Stock Prices
Latest ResNet and their variations (ResNeXt, Wide ResNet, PyramdNet) have broken the records of lowest error rates but they often suffer from problems such as vanishing gradients.
Read More


A Survival Analysis Method for Stock Market Prediction
The trading strategy based on the predictive regression model that includes trading information from both markets provides significant utility gains to mean-variance investors.
Read More


Intraday volatility interaction between the crude oil and equity markets
The trading strategy based on the predictive regression model that includes trading information from both markets provides significant utility gains to mean-variance investors.
Read More


Forecasting stock market short –term trends using a neuro-fuzzy based methodology
Paper proposes a new architecture that reformulate the layer inputs and provides comprehensive empirical evidence showing that these residual networks are easier to optimize.
Read More

Technical Architectures
Insights to quantitative methods applied to investment strategies, focused on machine learning and advanced computing.

Hyper-parameter Tuning
We have a set of hyper-parameters and we aim to find the right combination of their values which can help us to find either the minimum (eg. loss) or the maximum (eg. accuracy) of a function.
Read More

16-Dec-2019

Interpretable AI
Models can be rendered powerless unless they can be interpreted, and the process of human interpretation follows rules that go well beyond technical prowess.
Read More

16-Dec-2019

Convolutional Neural Networks
CNN replaces fully connected multiperceptron layers with convolutional layers and pooling layers so that the model requires a minimal number of parameters.
Read More


RLPG Report
Inspired by the idea of learning through interaction with environment, reinforcement learning (RL) is an area of machine learning concerned with what to do – how to map situations to actions in an environment so as to maximize some notion of cumulative reward.
Read More


Group Normalization
Group Normalization provides a more stable data normalization methodology that can ease optimization and enable very deep networks to converge.
Read More


ShakeDrop Regularization
Latest ResNet and their variations (ResNeXt, Wide ResNet, PyramdNet) have broken the records of lowest error rates but they often suffer from problems such as vanishing gradients.
Read More


Deep Residual Learning for Image Recognition
Paper proposes a new architecture that reformulate the layer inputs and provides comprehensive empirical evidence showing that these residual networks are easier to optimize.
Read More


Generative Adversarial Nets
GAN is at least competitive with the better generative models in the literature and highlight the potential of the adversarial framework.
Read More


Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
In Deep Neural Networks, inputs of each layers usually changes during training, which therefore requires lower learning rates and careful parameter initialization.
Read More


Deep Extreme Learning Machines
The emerging algorithm, Extreme Learning Machine, is known to be fast to train in comparison with iterative training methods, and performs with similar accuracy with Support Vector Machines(SVM).
Read More


Proximal Policy Optimization
The purpose of this research is to introduce a new family of policy gradient methods for reinforcement learning called PPO.
Read More


Ladder Variational Autoencoders
VAE, consisting of hierarchies of conditional stochastic variables, are highly expressive models retaining the computational efficiency of fully factorized models.
Read More


WHITE PAPER
Extreme Learning Machine
Algo Depth focuses much of its attention on signal processing, machine learning, and deep learning. In this extreme learning machine strategy, we apply a neural network to predict upper and lower bounds of future stock movements for the largest 25 companies traded on the Nasdaq exchange.
2017
Read More

WHITE PAPER
Application of Reinforcement Learning For Order Execution
We present a large scale application of reinforcement learning to optimize trading execution. The experiment is based on Apple Inc. The results generate a 1.2% annual cost savings, and show the promise of applying reinforcement learning methods to solve market microstructure problems. Our algorithm can further improve by including tick data, and market order books. It can be tested on any stock.