In this post, we discuss an approach to identify and analyze different market/correlation regimes extracted by clustering time-series data of the broad US ETF market. We present interesting visuals that analyze how markets in-general behaves in each cluster/regime, and use the insights obtained from the analysis to present a strategy selection/security selection model. The following sections detail the clustering methodology, results of time series clustering, (which includes broad market performance, analysis of correlation structure, network analysis for each cluster), and practical applications. The methodology used in this post is inspired by [1] to some extent.

Clustering Methodology

The majority of work done in classification of price moves using machine learning models use raw data even for an image classification algorithm like Convolutional Neural Network (CNN). However to highly dynamic and non-stationarity nature of stock prices these models either fail to capture any meaningful patterns or highly overfit to the training data (which is exacerbated for Deep Learning Models).

In this article we take a look at how state-of-the-art quantum machine learning classifiers perform against a classical machine learning algorithm like multi-layer perceptron and check if it exhibits any form of quantum advantage over the classical algorithm.

This article is a continuation of my previous post on using synthetic prices generated by MCMC to select strategy parameters. In this article, we discuss how one can augment the framework defined in the previous post to generate specific scenario based prices. Tactical Investment Algorithms Paper by Marcos Lopez de Prado 1, is one of the best papers around that details the power of using synthetic prices and MC backtests to find trading strategies optimized to work in a particular market regime i.e. trading tactically.

Note: The ideas and the backtesting methodologies detailed in this article are for illustrative purposes only, it's is in no way the best representation of how strategies must be backtested (same goes for metrics used to judge the strategy performance). The main purpose of this article is to exhibit how synthetic prices can augment/enhance the parameter selection process for a given strategy to reduce overfitting. This GitHub repository contains all the notebooks used in this article.