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 document introduces a Python library for adding technical analysis features to financial time series containing open, high, low, close, and volume data. It describes using the library with pandas and shows two workflows: adding a broad set of indicators…
The article describes an architecture that combines iTick market feeds, other economic and alternative data, and DeepSeek-R1 to support quantitative research workflows. It proposes attention-based data fusion, automated strategy variant generation,…
This research report explains reinforcement learning concepts, including Markov decision processes, state and action values, reward design, and value-based and policy-based algorithms. Its applied example uses a deep Q-network to time long positions in the…
This broad practical guide introduces machine learning in quantitative investing, covering supervised, unsupervised, and reinforcement learning. It explains data preparation, missing values, categorical features, scaling, train and test splits,…
The post proposes screening Chinese stocks in the metaverse theme using an opening price-change threshold and a candlestick reversal condition called a Morning Star. Its stated intent is to identify candidates for a potential oversold rebound. It also…
This assignment outlines a workflow for creating a visual machine learning stock strategy using BigQuant’s StockRanker course and strategy template. The participant is asked to modify the model’s input features by adding personal factors, then backtest the…
The document introduces convex functions through their geometric shape and explains a first-order characterization for differentiable convex functions on an open convex domain. At any point, the function’s value at another point is at least as large as the…
This document contrasts MQL’s built-in MathRand generator with a class wrapping a 32-bit PCG pseudorandom number generator. It characterizes MathRand as a limited linear congruential generator and says its statistical flaws make it unsuitable for large Monte…
This code outlines a daily trading workflow in which separate AI agents research a universe of leveraged exchange-traded funds, argue bullish and bearish cases, and pass their summaries to a trading judge. The universe includes leveraged long and inverse…
The document briefly describes an Expert Advisor that combines the Relative Strength Index with a small neural network. It says the system processes market data in OHLC mode during optimization and only acts when a new bar appears. An EURUSD five-minute…
This tutorial outlines a workflow for modeling Chinese steel futures using high frequency market data. It starts with tick snapshots for the RB2305 contract, explains that futures may trade during night sessions, and resamples the raw observations into ten…
This introductory assignment defines quantitative investing as using computer programs and historical market data to identify assets and apply predefined buying and selling rules. It highlights the ability to scan a broad universe more quickly than a human…
This document describes a competition built around Deep Alpha-CNN, an equity prediction model using a seven-layer one-dimensional convolutional neural network with residual connections. The report’s stated findings favor the deeper network over a two-layer…
This study summary describes a text-based factor intended to capture post-earnings-announcement drift in Chinese equities. It uses analyst report titles and abstracts about earnings forecasts, converts selected words into frequency features, and trains…
This document describes a demonstration of a backpropagation neural network (BPNN) for training and testing time-series predictions in MQL5. It says the implementation was ported from a C++ library and is presented as a coding demo rather than a real-world…
The article introduces least squares as a way to fit an empirical relationship to observed data. It illustrates a straight-line model relating ship length to width, first using two observations to determine a line and then explaining why that line need not…
This article introduces unsupervised clustering and contrasts it with supervised classification: cluster labels are not known in advance, so judging whether a grouping is correct is difficult. The goal is to group observations so that members of each cluster…
This guide describes a MetaTrader 5 script that creates a prompt file for customizing an AI assistant’s response language, persona, and numbered menu. Users can generate the file, apply it in the assistant, and edit the instructions, menu labels, and actions…
This guide explains why a submitted end-to-end Transformer example can run out of memory before training or inference begins: the data-loading step materializes minute bars for many stocks across the full date range in one DataFrame. Although the eventual…
This tutorial outlines a preprocessing workflow for equity factor data. It filters out records without an industry classification, fills missing factor values with date and industry group means, and describes clipping outliers using percentile, standard…
The article examines how training-window length can affect an AI stock-selection model. It compares longer historical samples with shorter windows using manual rolling training and a held-out backtest period. In the reported experiments, changing the…
This short discussion defines quantitative investing as analyzing financial-market signals with mathematical models to seek investment returns, with automated order placement as a possible part of the process. It identifies reduced emotional influence, the…
The document outlines a workflow for developing a simple EUR/USD trading system with market data from an API and code generation through an AI coding tool. Its example strategy buys when a short moving average crosses above a longer one and sells on the…
This article presents a supervised learning-to-rank approach to stock selection. It defines a universe using the constituents of the SSE 50, labels stocks by discretizing their subsequent five-day returns into 20 ordered classes, and trains on data from 2011…