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 configuration specifies a Qlib workflow for predicting CSI 300 stock returns with a gated recurrent unit (GRU) neural network. It uses Alpha158 features, filters to a selected feature set, applies robust normalization and missing-value filling, and…
This research summary explains why predictive machine-learning models can mistake unstable associations for causal relationships, especially when selection bias or confounding drives observed patterns. It presents propensity-score analysis as a way to make…
This lesson explains binary classification through logistic regression, using the example of predicting whether a stock will rise. It describes encoding outcomes as zero and one, interpreting the model output as a probability, and applying a 0.5 cutoff to…
This reflective account traces a trader’s attempts to find reliable market predictions. The author moves from familiar indicators such as MACD, KDJ, and RSI to custom indicator programming, time-series methods including ARIMA and GARCH, neural networks,…
The document explains Qlib’s meta-controller framework for learning patterns across forecasting tasks and using them to guide future tasks. A meta-task packages data for a meta-model, while a meta-dataset manages how task information is generated and…
This overview explains how quantitative strategies use factors to express hypotheses about future returns. It describes candidate signal sources such as prices, fundamentals, macroeconomic data, and text, then outlines a research process: translate an…
A BigQuant support discussion addresses a user’s difficulty training a StockRanker model for individual-stock prediction. The replies recommend moving to the newer StockRanker template because the older module is no longer actively maintained. A second…
This brief platform discussion asks how to inspect the output of an intermediate neural-network layer in a visual workflow. One reply recommends adding and running a notebook cell after execution, inspecting the layer object’s properties, and reading its…
The document describes a daily directional forecasting strategy for SPY using a quadratic discriminant analysis model trained on the prior two S&P 500 returns. It maps predicted up days to a fixed long position and predicted down days to a fixed short…
The document introduces matrix inversion as a way to solve systems of linear equations, connecting the idea to applications that include quantitative finance, statistics, and machine learning. It explains that the identity matrix leaves a matrix unchanged…
This brief overview proposes a two-stage automated equity approach. Machine learning would rank or filter stocks for potential, while reinforcement learning would determine when to buy or sell the selected names. The intended result is a system that connects…
This platform support thread discusses how to preserve a model after rolling training and whether the final saved artifact represents the intended result. The replies characterize rolling training as a research-stage feature rather than a direct route to…
The article describes turning rolling OHLC windows into sparse, binary candlestick images for a convolutional neural network. Each sample uses 32 time steps, normalized and mapped onto a 128-by-128 grid; three columns represent each candle’s open, high-low…
This configuration defines a Qlib equity-prediction experiment for China’s CSI 500 universe. It trains a double-ensemble model built on gradient boosting, using Alpha360 features and a forward close-to-close return label. Training and validation segments…
This report summary describes revisions to AlphaNet, a neural network that learns stock-selection factors directly from raw price and volume data. AlphaNet v2 adds six ratio features, replaces pooling and dense layers with an LSTM to capture temporal…
This introductory overview explains ten general-purpose algorithms and problem-solving methods: quicksort, heapsort, merge sort, binary search, BFPRT selection, depth-first and breadth-first search, Dijkstra’s shortest paths, dynamic programming, and naive…
The document outlines an AI-assisted workflow for turning natural-language strategy ideas into code and backtests. It distinguishes Python strategies, suited to broader stock selection, multi-factor models, portfolio rules, and risk controls, from formula…
The document reviews tactical and dynamic asset allocation, then focuses on predicting short-term movements in Chinese equity indices. It compares logistic regression, artificial neural networks, and support vector machines across different inputs and…
This page introduces Andrew Ng’s machine learning course and lists its main subject areas and accompanying study materials. It describes supervised learning methods such as neural networks and support vector machines, alongside unsupervised learning topics…
This announcement outlines a two-day VeighNa training course on machine-learning cross-sectional multi-factor strategies, also described as alpha strategies. It presents their use in equity selection, index enhancement, absolute-return portfolios, and…
The document introduces hyperparameter optimization for machine learning models used in trading strategies. It describes how choices such as a model's learning rate can change the path taken while minimizing a loss function, which may have multiple local…
This overview explains why high-frequency research systems process tick snapshots, trades, and order data with distributed computing. It describes the challenge of loading very large datasets and the time required to clean and analyze them on a single…
This example shows how to assemble a quantitative equity research workflow in Python using Qlib. It initializes Chinese market data, creates a model and dataset from a task configuration, fits the model, and records predictions and signal analysis. It then…
This tutorial outlines a paper trading workflow for a FinRL stock trading agent. It begins with installing the library and preparing Alpaca paper account credentials, then introduces a Proximal Policy Optimization agent with actor and critic networks. The…