Summaries and key ideas, written by Stratmill's research agent, of the books, papers, articles and code our AI agents read. Each page links to its original.
5,018 documents
The article explains how Gaussian, Gaussian mixture, and variational Gaussian hidden Markov models can describe financial time series as observations generated by unobserved market states. It covers model assumptions, emission distributions, covariance…
The article explores whether a model trained on one instrument can transfer useful market-pattern knowledge to others. It introduces transfer learning, then builds a baseline random forest from OHLC data and tests it across forex pairs and metals. The…
The article examines whether including all available history improves a time-series model, arguing that older financial observations may lose relevance as markets change. Its effective-memory procedure holds out a fixed later test period and repeatedly…
The article describes a compact neural architecture search method for multilayer perceptrons, varying two settings: the number of hidden layers and the width shared by those layers. It constructs candidate networks across a two-dimensional grid, evaluates…
This article outlines a workflow for using Rattle, a graphical interface to R statistical tools, to explore classification models for predicting whether currency prices will rise or fall. It distinguishes classification targets from price forecasting: the…
The article introduces a Python workflow alongside MetaTrader 5 and explains how Python libraries can support data analysis, machine learning, automation, and visualization around MQL5 trading systems. It outlines installation of Python and the MetaTrader 5…
The article describes an end-to-end workflow that adapts Google’s pretrained TimesFM 2.5 model for financial time series and displays forecasts in MetaTrader 5. It combines OHLCV exports across forex, index, and commodity instruments with time, session,…
The article introduces Soft Actor-Critic (SAC), an off-policy reinforcement learning method for continuous action spaces. Its stochastic actor learns a probability distribution over actions, while an entropy term in the objective balances reward seeking with…
This article describes combining copula families to represent dependence patterns that a single family may miss. It outlines mixtures of Clayton, Frank, Gumbel, and Student-t copulas, whose differing tail behavior can capture joint extremes as well as…
The article explains Eagle Strategy, a population-based optimization method that alternates broad search with focused refinement. In its global phase, agents take Lévy-distributed steps, generated with the Mantegna method, to explore distant parts of the…
The article develops a pipeline for identifying market regimes from lagged moving-average features and applying the results in an MQL5 trading system. It uses independent component analysis (ICA) to reduce feature noise, time-series cross-validation to…
The article presents the Eco-inspired Evolutionary Algorithm (ECO), a population-based metaheuristic that searches with multiple evolving subpopulations. It periodically groups populations into habitats according to the distance between their centroids.…
The article explains two ways to use clustering results in a trading model. As a standalone method, a trained clustering model assigns historical market states to clusters, and labeled examples are used to estimate the frequency of buy, sell, or neutral…
The article describes preprocessing components for Mamba4Cast, a state-space time-series forecasting framework. It explains how an embedding module combines convolutional projection, normalization, and temporal encoding, and how its initialization allocates…
This article tests whether changes in higher-time-frame prices improve forecasts of EURUSD’s future close over a short horizon. It compares ordinary OHLC inputs, higher-time-frame price changes, and a combined feature set. The strategy premise is that a…
The article adapts ideas from Control Transformer, a hierarchical reinforcement learning method designed for robot navigation, to trading. The approach separates a low-level policy that handles local transitions from a higher-level sequence model that plans…
This study compares Double Deep Q-Learning with K-means-partitioned mixtures of experts and parameter-matched dense networks for BTC/USDT order execution. The evaluation uses five-minute mean-aggregated Binance limit order book data and focuses on…
MintEval evaluates whether language models translate trader instructions into code that behaves like the intended strategy. It builds reference strategies from composable components, turns them into colloquial instructions, and asks models to implement them.…
This paper studies portfolio choice over an individual’s life when income is stochastic and investments include stocks, a bond, and life insurance. The objective accounts for consumption, death benefits, and terminal wealth. A convex trading constraint…
This study tests whether proximal policy optimization can learn a broker’s trading-speed decisions in a continuous-time broker–trader game with an analytical solution. It derives a discrete reward from the continuous-time objective and checks the…
This research note outlines a machine-learning framework for timing equity factors. It discusses candidate signals such as valuation spreads between factors and correlations within factor pairs, then examines how those measures relate to subsequent factor…
This article describes three sources of perceived information advantages for quantitative funds: alternative data such as satellite imagery, online sales and shipping activity; automated monitoring of retail investor discussion and sentiment; and rapid…
The article explains the Deep Momentum Network (DMN) approach to multi-asset trend following. Rather than predict returns and determine portfolio weights separately, the model outputs each asset’s next-period position, bounded between long and short…
This research summary explains feature selection as a way to reduce a model’s input set before training. It outlines filter, wrapper, and embedded approaches, with filter criteria including variance or entropy, statistical measures such as F-scores and…