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 summarizes CapTE, a model for predicting stock movements from social media text. A Transformer encoder extracts semantic features from posts, while a capsule network is used to represent structural relationships in the text. The approach is…
The document introduces Temporal Routing Adaptor (TRA), a model designed to learn multiple trading patterns from stock market data. It describes using TRA with Qlib datasets and workflows, and notes that the paper’s reproduction setup first trains a backbone…
The document summarizes a study that develops a probabilistic classifier to identify high-frequency trading activity from intraday order data. Using French BEDOFIH market records, the researchers engineered features describing orders, including their prices,…
The document summary highlights two applications of machine learning in quantitative investing. First, it describes forecasting volatility to inform how capital is allocated among strategies, based on the claim that many strategies’ profitability is closely…
The document explains how Qlib’s tuner searches hyperparameters and combinations of models, trainers, strategies, and data labels. A configuration defines each tuner’s search spaces and evaluation limit, then organizes tuners into a pipeline. Users choose a…
This beginner tutorial uses the MNIST handwritten digit dataset to introduce TensorFlow through a simple image classification task. Each image has a digit label, and the model is intended to predict that label from the image. The tutorial chooses softmax…
This article challenges three barriers commonly associated with quantitative investing: needing advanced mathematical credentials, being able to code extensively, and having a large portfolio. It presents quantitative analysis as a way to use statistics and…
The report outlines a framework for timing equity factors whose performance has become less stable. It first examines indicators such as valuation spreads and pairwise correlations, testing their relationship with future factor returns. It then uses a random…
This presentation interprets findings from a 2021 survey of Chinese quantitative investment institutions and discusses how the sector was developing at that time. It covers strategy mixes, research organization, talent, artificial intelligence, alternative…
This meetup Q&A contrasts futures CTA strategies, often framed around trend following, with equity multi-factor strategies that combine signals such as value, momentum, quality, and size. It outlines a Bollinger Band example for futures: calculate a…
The document introduces VeighNa, an open-source Python framework for quantitative trading, with particular attention to its vnpy.alpha module. That module organizes research into feature creation, model training, strategy development, and workflow…
This tutorial demonstrates a graph convolutional policy, GPM, inside a reinforcement-learning portfolio workflow. It loads historical stock features and a sector and industry graph, then reduces the graph to nodes within two hops of the selected portfolio…
This brief description introduces a reinforcement-learning lecture on the theoretical foundations of dynamic programming. It says the lecture studies dynamic-programming algorithms as contraction mappings and asks when and how those mappings converge to the…
This paper summary examines whether historical trading data can predict the next month’s cross-sectional returns of Chinese A-shares. It describes a dataset of 108 stock characteristics from 1997 to 2019 and compares traditional econometric methods with six…
The document presents daily portfolio rebalancing as a Markov decision process. An agent selects nonnegative weights for Dow 30 stocks, normalized to sum to one, using a state that combines a rolling covariance matrix with MACD, RSI, CCI, and ADX indicators.…
This article introduces XGBoost as a machine-learning method for quantitative stock selection using price and volume factors. It explains boosting as a process that adds weak learners in sequence, then contrasts AdaBoost’s reweighting of misclassified…
The document raises a portfolio-construction question about using a stock-ranking model to select both ends of its predictions: stocks with the highest factor scores and stocks with the lowest scores. The proposed idea is to hold the two groups together as a…
This article introduces support vector machines for classification and regression, then applies them to A-share stock selection. It explains the maximum-margin principle for linear SVMs, slack variables for imperfectly separable observations, and kernel…
This tutorial explains how support vector machines classify data by finding a boundary with a wide margin, and how slack variables allow some classification errors in noisy data. It introduces kernel methods as a way to handle nonlinear boundaries by…
This configuration describes a Qlib experiment using a graph attention model, GATs, with an LSTM base model to predict near-term returns for CSI 300 constituents. It sets Chinese market data, defines a close-to-close forward return label, normalizes features…
This short Chinese-language note defines quantitative investing as using programs to invest based on collecting and analyzing substantial market data. It presents automation as a way to respond to market changes more quickly, follow a consistent process, and…
This tutorial explains applying principal component analysis to stock returns to identify dominant co-movement patterns. It standardizes historical returns, estimates a rolling correlation matrix, and decomposes it into eigenvalues and eigenvectors. The…
This Chinese-language forum post concerns calculating higher-timeframe indicators in real time from a lower-timeframe bar callback, such as updating 30- or 60-minute KDJ or RSI while processing five-minute bars. The example creates separate bar generators…
The document contrasts an educational AI investing project, which organizes investor-style agents to debate ideas, with a framework centered on the trading strategy lifecycle. It describes a workflow in which agent decisions are tested on historical data,…