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
This overview surveys several ways to rank or select equities using quantitative models. It groups approaches into factor and risk models, statistical methods, machine learning, macroeconomic inputs, and behavioral signals. Examples include combining value,…
The article introduces random forests as an ensemble of decision trees and explains how bootstrap sampling and random feature selection create diversity among trees. It reviews classification and regression trees, impurity measures, pruning, regression…
This report summary presents findings from interviews with 100 senior professionals at asset managers, wealth managers, hedge funds, and digital asset firms. It describes institutional concerns about fragmented regulation, interoperability for tokenized…
This tutorial excerpt introduces a data-preparation workflow for FinRL. It describes downloading historical equity prices in open, high, low, close, and volume form, and notes that FinRL’s Yahoo downloader uses adjusted closing prices and adds a weekday…
The research constructs a specific market-capitalization factor by modeling cross-sectional stock market value from company financial measures and market variables. The model’s residual—actual market capitalization relative to the value estimated from those…
This overview introduces Python libraries used across quantitative trading workflows. It groups tools by task: NumPy for numerical operations, Pandas for structured and time-series data, TA-Lib for technical indicators, and Matplotlib or Plotly for…
A platform forum exchange addresses whether a daily backtest that predicts a five-day return using the current day’s closing price can access that day’s prediction before the close is available. The response explains that, in the described daily setup, the…
This guide presents local checks for two common failures in factor submissions: missing values near the evaluation window’s start and look-ahead bias. For time-series factors, it recommends querying enough earlier data to cover the longest lag or rolling…
The document discusses an interface pattern for chart-based research: select a range of candlesticks, then ask a question about that specific interval. A selection gives an otherwise vague reference such as a question about why a move occurred a defined…
This Qlib example shows an online manager coordinating rolling model training and strategy updates. Its workflow begins by training an initial strategy set, then runs a routine that updates online predictions, prepares new tasks and models, and generates…
The guide lays out different entry paths for using QlibRL, depending on whether the reader is new to reinforcement learning, researches RL algorithms, or already has quantitative finance experience. It recommends learning RL fundamentals, understanding…
This Chinese-language summary presents a 2021 survey of reinforcement learning in financial decision-making. It introduces Markov decision processes as a foundation, then identifies value-based and policy-based learning approaches, including extensions that…
This strategy combines model-based stock ranking with technical timing. It starts from a manually defined CSI 150-related universe, then ranks candidates using an AI prediction score and volume, turnover, and large-order money-flow factors. Timing filters…
This guide surveys how machine learning can support quantitative investing, from data preparation and model evaluation to portfolio selection. It explains supervised, unsupervised, and reinforcement learning, and reviews methods including regularized…
This Chinese strategy discussion combines hand-built stock selection rules with a CNN classifier. It describes converting screening conditions into Boolean features so a model can learn to identify stocks expected to show short-term strength. The feature…
The article explains genetic algorithms as population-based search methods inspired by biological evolution. It walks through encoding candidate solutions as chromosomes, evaluating them with a fitness function, selecting parents, combining them through…
This study applies Lasso, adaptive Lasso, and Elastic Net to select equity factors and combine them into stock-selection signals. Its motivation is that correlated predictors can make ordinary multivariate regression unstable. The reported experiments show…
The document summarizes research on improving a primary sequence-learning task by selecting among related tasks that differ in timing. Examples include translation with different input delays and stock trend prediction over different future horizons. Because…
This configuration describes a Qlib data pipeline for five-minute CSI 300 constituent data over a specified two-year period. It defines training, validation, and test segments, with the training interval also used to fit feature processing. The feature…
The document explains feature engineering as the process of turning raw data into variables that better represent a prediction task. It argues that useful features matter in quantitative strategy research because market data and factor choices constrain what…
The document describes a hybrid trading system in which an ONNX model classifies market structure and conventional rules govern trade selection and management. Model inputs include ATR-scaled momentum, distances from recent swing extremes, relative tick…
The document raises a label-design question for equity strategy development. It observes that labels based only on shifted closing and opening prices may not distinguish stocks that rise from those that fall, and asks whether adding highs, lows, and volume…
This article proposes grouping a StockRanker model's scores instead of assuming that the highest-ranked stock is consistently the best choice. It notes that which rank performs well can vary from day to day, especially for models whose styles are unstable.…
This brief BigQuant support exchange explains a feature-name mismatch that prevents a LightGBM module from locating an input column. The module is given a calculated expression as the feature name, while the DataFrame column is actually named “ma.” The…