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 answer explains conditional likelihood as a way to estimate models in which the distribution of observations changes over time. In a standard maximum likelihood setup, the same parameter values govern each observation. With time-varying parameters, each…
The document asks which unsupervised learning methods can help discover patterns in manually selected financial time-series features, with the aim of making broad predictions about future price behavior. It gives examples ranging from predicting the next…
The document presents a Python and QuantLib problem: a dividend-paying American option returns zero for its net present value and Greeks despite inputs that should produce nonzero values. The code builds an American exercise option, supplies discrete…
The document discusses how to align an expert-based internal scorecard for commercial loans with external agency rating categories. It suggests ordinal logistic regression when agency rating category is the target, since ratings are ordered categories rather…
The discussion questions whether reinforcement learning is useful when a portfolio model is trained on fixed historical prices. In that setup, market states advance independently of the chosen action, and next-period rewards can be deterministic given the…
This thesis-related question asks how to choose bounds for the Heston model parameters kappa, theta, sigma, and rho when generating synthetic data to train a neural network for calibration. It compares one published set of ranges, described as excluding…
The document considers forecasting actual high-frequency prices, including bid and ask quotes, rather than forecasting volatility. It highlights practical challenges: model validity may be brief, validity is difficult to assess, and computationally expensive…
The discussion distinguishes PCA scores, which describe observations such as dates in principal-component space, from factor loadings, which describe each stock’s exposure to the components. It explains why clustering loadings may emphasize a component with…
The document asks whether practitioners commonly test the state and measurement noise assumptions in a linear Kalman filter, including whether the noise should be Gaussian. Its answer distinguishes the filter’s estimation objective from the stronger…
The document considers how to assess trading-signal performance across instruments, sectors, volatility levels, and combinations of market conditions, while managing a large number of possible segments. One response warns that increasingly fine-grained…
The discussion separates the computational cost of option pricing from the difficulty of modeling market values. Under Black–Scholes–Merton assumptions, a European option has a closed-form price with little computational work. American options require…
The discussion surveys simple ways to project a company’s earnings from historical monthly data. Suggested starting points include plotting the series to inspect its behavior, fitting a log-linear trend, or applying an autoregressive model to a stationary…
The document explains purging in financial machine learning through a prediction task that uses current stock returns to forecast the next day’s market return. The accepted answer illustrates the issue with a label based on a multi-day future return: labels…
The document considers rolling Shannon entropy of intraday returns and asks how histogram bin size affects comparisons over time. Its answer suggests mutual information when the goal is to measure dependence between variables, with normalization by its…
The document asks how to choose stop-loss and take-profit thresholds from trade outcomes and expected profit estimates. It points to operations research and stochastic processes, and presents differing views: one answer argues that thresholds may be…
The document considers how to choose both a signal threshold and the future return horizon when evaluating a trading hypothesis. It warns that searching many combinations can amount to data mining and produce overfit results, while acknowledging that there…
This document describes a simulation problem: generating several correlated variables from a Sobol quasi-random sequence in MATLAB. The author contrasts a Cholesky-based transformation of standard normal draws with an attempted method that transforms Sobol…
The document describes a simulation in which Poisson counts determine how many exponential random variables to draw for each component, after which weighted component sums produce one value per path. The original example uses nested Python loops and targets…
The document asks why securitized-product prepayment models often represent borrower status with discrete delinquency states and estimate transition probabilities, rather than directly regressing the dollar amount prepaid in each period. It notes that…
The discussion considers how to obtain timestamped news for research linking Forex price series to news releases, especially when professional machine-readable feeds are costly. Suggestions include collecting headlines from websites, parsing RSS or Atom…
The document asks whether quants build one prediction model per stock or train a model that can generalize across many securities. It highlights two concerns with training on a single stock and applying that model elsewhere: survivorship bias and the narrow…
The document addresses how to find long historical series of nonfinancial measurements for testing correlations with a financial analysis or trading algorithm. It suggests looking at unusual public datasets and gives examples of possible sources, including…
The document discusses a two-stage classification approach to finding trading opportunities. A primary model is tuned for high recall, accepting false positives so that it identifies most potential positive cases. A secondary model then attempts to filter…
The document separates machine learning used to approximate an already computed option pricing function from methods that learn both a claim’s value and its optimal hedge. For fast pricing, a model is trained on examples generated by slower numerical pricing…