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 Qlib documentation explains its experiment-management structure for organizing quantitative research. An experiment manager oversees experiments, each experiment groups recorders, and each recorder represents an individual run. The high-level Qlib…
This document outlines an asset-allocation approach that separates asset selection from portfolio weighting and market timing. It explains a hidden Markov model as a way to infer unobserved market regimes from observable data, represent complex observations…
This brief troubleshooting exchange addresses a quant model that fails when configured with a single factor, even though the factor values appear normal. The proposed fix is to assign the factor a shorter alias; the author reports that using a very brief…
This Qlib configuration defines a supervised stock-ranking experiment using the CSI 300 universe and the Sandwich neural model. Its data handler applies robust feature normalization with outlier clipping, fills missing feature values, drops missing labels,…
This overview introduces SciPy as a Python library built on NumPy for scientific and numerical computing. It surveys modules used for linear algebra, optimization, integration, interpolation, probability and statistical analysis, signal processing, and…
This brief BigQuant support note addresses a log message saying that a factor cannot find its dependent data. The reported issue is not missing factor metadata: the expression appears to place two factor names, close and average turnover, next to each other…
This article explains how residual networks address degradation in very deep neural networks. In a plain network, adding layers can worsen accuracy even when gradients are managed through initialization and other techniques. A residual block instead learns a…
This overview introduces the Transformer architecture from the paper “Attention Is All You Need.” Earlier sequence-to-sequence systems commonly used recurrent or convolutional networks, often combined with attention. The Transformer instead relies on…
This overview introduces the Transformer architecture from the paper “Attention Is All You Need.” Earlier sequence-to-sequence systems commonly used recurrent or convolutional networks, often combined with attention. The Transformer instead relies on…
The post documents an attempt to move daily Chinese stock bars into a custom data source and use them as input to an automated labeling module. Its example queries a date-bounded range of stock records, adjusts instrument identifiers, writes the resulting…
This Qlib workflow example combines an LightGBM model trained on Alpha158 features with a top-ranked stock strategy for the CSI 300 universe. It sets training, validation, and test periods, then demonstrates nested execution across daily, 30-minute, and…
The document explains Lumibot’s local memory system for AI trading agents. It uses SQLite to keep an append-only event history, searchable current views of memories and theses, and records of what the agent retrieved. Parquet exports support later review and…
This document summarizes a 2012 study that trained a deep convolutional neural network to classify high-resolution images into 1,000 categories. The model used convolutional and fully connected layers, GPU-based computation to speed training, and dropout to…
The document raises a data construction question for a rolling-trained deep learning model that predicts a stock’s next-day outcome from a fixed-length history. The dataset concatenates each stock’s chronological observations, then randomly selects training…
This article describes an A-share stock selection approach that uses XGBoost to combine Barra style factors, then examines whether selecting a smaller set of factors according to market conditions can improve the strategy. It compares three selection…
The document briefly explains how to inspect a trained StockRanker model in BigQuant's strategy generator. After training, the platform can display feature importance and the model's tree branch structure. These views can help users see which inputs…
The document presents a market-timing experiment that applies a Gaussian hidden Markov model to daily CSI 300 data. Its feature set combines the one-day log return, five-day log return, and log high-low range. The model is configured with six latent states…
This introductory note defines quantitative investing as using numerical methods, models, and algorithms to make investment decisions, including automated execution. It lists potential advantages: rules can reduce emotion-driven choices, computers can…
The document presents a QuantStats tear sheet for a strategy labeled tqqq-plain, compared with SPY. It reports a short backtest covering January 4–15, 2026, using Yahoo data, alongside return, drawdown, volatility, risk-adjusted performance, and benchmark…
The report introduces unsupervised learning, focusing on dimensionality reduction and clustering for financial analysis. It describes factor analysis and principal component analysis as ways to compress many stock or bond characteristics into a smaller set…
This document summarizes a machine learning timing strategy that uses a long short-term memory network to model stock market time series. It presents the model as a way to learn dependencies across observations and adapt to sharp changes in a trend. The…
This research summary outlines a progression from single-factor IC and IR analysis toward attribution in machine-learning models and causal analysis. It says IC and IR are foundational for traditional multifactor research but do not capture interactions…
The AI Trend Navigator is a ProRealTime indicator that applies a nearest-neighbor approach to price-derived series. It compares a current target value with values in a historical window, selects the closest matches, and averages them into a smoothed…
This Chinese-language Q&A covers practical questions about equity factors, model templates, backtest assessment, and trading tools. It names market capitalization, turnover, momentum, quality, valuation, growth, and profitability as commonly used factors,…