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
QlibRL organizes a reinforcement learning workflow for trading around an environment wrapper that connects a policy to a market simulator. The wrapper accepts actions, advances the simulated market, and returns updated states and rewards. Its components…
This implementation builds a committee of neural network regressors, trains each member on the same training data with validation data and early stopping, then averages their predictions. The model class and parameters, committee size, training epochs, and…
This introductory overview explains ten practical principles of machine learning. It emphasizes that data quality and representativeness matter more than algorithmic complexity, and recommends simpler models when data is limited to reduce overfitting. It…
The article describes inferring a fund’s constituents and weights from its total value time series. It compares Extended Linear Cloning (ELC), which uses regression coefficients as asset importance scores and retains the highest scoring candidates, with…
This code utility builds pairwise dependence matrices from columns in a feature DataFrame. It supports information-based measures, distance correlation, rank correlation, GPR and GNPR distances, and optimal-transport dependence. Parameters let users…
This overview explains how AI quantitative investment platforms combine machine learning, deep learning, and natural language processing with financial data analysis. It describes a general workflow in which platforms collect and process market and financial…
This configuration specifies a Qlib workflow for training a LightGBM model on the Alpha158 feature set for China’s CSI 300 universe. It defines training, validation, and test periods, along with model settings such as mean squared error loss, tree depth,…
This overview introduces statistical learning methods for prediction and data analysis. It covers linear and logistic regression, linear and quadratic discriminant analysis, resampling, predictor selection, regularization, dimension reduction, nonlinear…
The article describes a long-only Chinese equity strategy that uses a ResNet101 image classifier to identify strong stocks during pullbacks in an established uptrend. Its rationale combines buying a retracement in a strong trend, favoring stocks that…
This tutorial introduces feedforward neural networks, convolutional neural networks, recurrent neural networks, and LSTMs. It explains how perceptrons combine features with learned weights and an activation function, how layers build representations, and how…
The article introduces financial time series as sequential observations modeled as realizations of stochastic processes. It identifies trends, seasonal variation, and serial dependence as recurring features, with volatility clustering and commodity…
The response explains how to reference the current close alongside earlier close values in a BigQuant factor expression. It constructs a series of conditional comparisons: for each of several prior indexed closes, it checks whether the current close is lower…
The document explains factors as measurable characteristics of assets that can serve as inputs to quantitative models and investment strategies. It outlines four common feature-engineering steps: cleaning missing or erroneous observations, standardizing…
The report combines cross-sectional equity factors with a set of three macro cycle indicators in a random forest model. The cycle features act as market state signals, allowing the model to apply different stock-selection logic across conditions. The authors…
This example describes an AI-assisted value-investing workflow inspired by Warren Buffett’s public approach. One agent reviews filings and assesses business quality, cash generation, balance-sheet strength, and durability. A second challenges the valuation…
This research summary presents a framework for replicating the returns of an index, fund, or individual stock with a selected set of holdings. It first applies LASSO regression to choose securities that capture the target portfolio’s characteristics, then…
This stock selection approach combines three filters: a price-range threshold, a bullish crossover of the five-week and ten-week moving averages, and a minimum time since listing. The post explains the crossover and trading-range conditions as technical…
This guide describes participation in a Chinese A-share quantitative challenge with factor discovery and end-to-end modeling tracks. Both target future 30-minute VWAP returns using minute bars and order book snapshots. For the factor track, submissions…
This brief troubleshooting exchange addresses why two backtests can report different returns even when their factors and backtest dates match. The response attributes the mismatch primarily to a difference in training-set data volume. It adds that using the…
This brief forum exchange addresses a feature calculation that displayed missing values and offers a warning relevant to strategy construction: using a shift with a negative offset can incorporate future observations into a feature. Such a feature may appear…
This brief response defines quantitative investing as using tools to extract patterns from historical data. It frames time-series analysis as a way to improve the odds of making a successful decision and cross-sectional analysis as a way to seek better…
The document explains a model-training error caused by a feature stored as an object or string rather than a numeric value. Its diagnostic approach is to inspect the training data and identify which input column has the incompatible type. The example points…
The document explains two settings in a rolling-data module used to prepare features for securities. The window size controls how many consecutive observations are grouped into each sample. Its example uses five sequential values and a window of three to…
This data-preparation workflow builds model inputs for a momentum strategy from asset closing prices. It clips prices using bounds based on an exponentially weighted mean and standard deviation, derives daily returns and volatility, and creates a next-period…