FinAcademy › Risk Track I › Simulation & ML

Module R7 · QTA 13–15

Simulation and learning machines

Monte Carlo and its standard error, variance reduction, the statistical bootstrap, then the two machine-learning chapters: data preparation, PCA and K-means, cross-validation, regularisation, logit, trees and ensembles, KNN, SVM, neural networks, reinforcement learning and NLP.

3 lessons · 8 interactive figures · 22 exercises

1. 1Monte Carlo, variance reduction and the resampling bootstrap
   Simulation error falls with the square root of the number of paths.7 exercises
2. 2Machine learning: data, PCA, K-means and validation
   Unsupervised learning finds patterns, supervised learning predicts, reinforcement learning decides.7 exercises
3. 3Prediction: logit, trees, ensembles, KNN, SVM, neural nets
   Default prediction is a classification problem. The exam wants the mechanics of each method and how to score it.8 exercises