@Demo public class Infantino2010PredictionAnalysisDemo extends Object
Using this example. Compares predicted and realized returns for the Infantino model and writes correlation analysis to log/pca-corr.csv.zip. This is forecast analysis rather than a funded portfolio simulation; a prediction correlation is not an account return.
Model and publication. PCA supplies a low-dimensional systematic-return representation for an accumulated-return regression. This example studies the forecasts on daily data, unlike the thesis’s intraday experiment.
Implementation in this tutorial. DynamicInfantino2010PCA receives T=500, k=2 and H=120 by default; N follows setupStocks(). MovingWindowOfReturns creates realized log-return comparisons and Covariance accumulates prediction/realization diagnostics. The CSV is a forecast-diagnostic output, not simulated account NAV. Overlapping H-period observations can be serially dependent; correlation alone does not establish tradable performance.
| Constructor and Description |
|---|
Infantino2010PredictionAnalysisDemo() |
| Modifier and Type | Method and Description |
|---|---|
static void |
main(String[] args) |
void |
run(List<Stock> stocks,
dev.nm.misc.datastructure.time.LocalDateTimeInterval interval) |
static List<Stock> |
setupStocks() |
public static void main(String[] args) throws dev.nm.misc.parallel.MultipleExecutionException, IOException
dev.nm.misc.parallel.MultipleExecutionExceptionIOExceptionpublic void run(List<Stock> stocks, dev.nm.misc.datastructure.time.LocalDateTimeInterval interval) throws dev.nm.misc.parallel.MultipleExecutionException, IOException
dev.nm.misc.parallel.MultipleExecutionExceptionIOExceptionCopyright © 2026. All rights reserved.