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 frames a thesis question about established, pre-neural-network approaches to calibrating stochastic volatility models, with particular interest in Heston. It asks for research areas and papers that could support a historical overview of…
The document asks how to forecast the next stock move after fitting and evaluating a hidden Markov model. It points to the transition matrix as a possible part of the answer, but does not explain how to use it or provide an actual forecasting procedure. Its…
This document surveys approaches to grouping time series that look similar, including curves with similar means but different variance or shape. One suggestion is to represent each series with informative features such as higher distribution moments,…
The document discusses adding candlestick-pattern features to a machine-learning model that predicts whether a stock’s next daily close will rise or fall. The researcher scores patterns by the share of observed cases in which price moved in the pattern’s…
The document addresses a mistaken either-or choice between modeling returns with a parametric probability distribution and using an autoregressive model. These describe different parts of a model: a distribution specifies the form of uncertainty, while…
The document raises practical questions about using XGBoost to predict whether the next candle’s closing price will move above or below a chosen threshold from the current open. The proposed feature set combines market sentiment, macroeconomic variables,…
The document asks how to adapt the differential Sharpe ratio reward used in a trading reinforcement learning system so that it instead reflects the Sortino ratio. The proposed motivation is that the traded assets are volatile and positively skewed, making…
The document offers research guidance for studying whether neural networks can predict liquidity or bid-ask spreads from market depth data. It recommends preparing the data carefully, considering zero or negative spreads, and planning how limited…
The document frames look-ahead leakage checks as mutation testing. At a historical cutoff, a valid mutation preserves information available at that time while changing only future data; if the historical artifact changes, this suggests the pipeline used…
The document examines a beginner’s attempt to use a fast Fourier transform (FFT) on Apple’s adjusted closing prices, including plots reconstructed with only a few frequency components. The resulting curves fall sharply near the end of the sample, prompting a…
The document asks how to model prices that are pulled toward a target value but may overshoot it when momentum persists. It presents a discrete return regression with both a lagged-return term and a target-gap term, then reviews possible continuous-time…
The document discusses preparing technical indicator time series as inputs to a neural network. It presents two basic transformations: use differences between consecutive observations, or use percentage changes to express relative movement. When combining…
The document considers a daily classification task: using index features to predict whether the next day's open-to-close move is up or down. The proposed workflow forms a direction label, selects and scales features, trains an SVM on a rolling window, and…
The question considers forecasting a future multi-day return from a history of rolling returns. Because adjacent rolling windows share most of their daily observations, the inputs are strongly autocorrelated. The response says that autocorrelation alone does…
The document discusses potential drawbacks of using deep learning to calibrate and price a full implied-volatility surface. It notes that neural methods can produce very fast calibration at application time, while training may involve substantial…
The discussion compares graphical machine-learning environments such as RapidMiner, KNIME, and WEKA with programming tools including R, Python, and scikit-learn. It frames the choice around platform familiarity, learning time, dataset size, need for…
The document collects public claims and opinions about Renaissance Technologies’ Medallion fund rather than establishing a definitive account of its trading strategy. Suggested techniques include algorithmic trading, information theory, hidden Markov models,…
The document describes a research need for reliable Heston option prices, especially for options far out of the money. The author reports that numerical instability in a current implementation produces slightly negative or oscillating prices, which then…
The document describes a proposed research approach: define learning environments from market microstructure models with known solutions, then use reinforcement learning to see whether agents recover features of those equilibria. As an initial example, the…
The document explains how least-squares Monte Carlo (LSM) values an American option and identifies when exercising is preferable to continuing to hold it. Working backward from terminal payoff, the method estimates each time step’s discounted continuation…
The document considers how to use rapidly changing limit order book features, such as spread and bid-ask imbalance, to predict a future midprice. It suggests testing recent history or aggregation and emphasizes that useful inputs may include time of day,…
The document surveys approaches to estimating time-varying dependence with copulas in R. It points to vine copulas with evolving parameters as one research direction and mentions the CDVine package and related materials for statistical inference. It also…
The document raises two questions about reproducing a hidden Markov model study of factor investing across market regimes. First, it asks whether the reported regime labels and their return and volatility summaries use future observations to classify past…
The document addresses how traders might identify market conditions associated with changes in algorithmic strategy performance. It proposes structural-break analysis as one avenue, naming CUSUM tests on recursive residuals or levels and explosiveness tests…