The University of Southampton
University of Southampton Institutional Repository

Constructing smart financial portfolios from data driven quantitative investment models

Constructing smart financial portfolios from data driven quantitative investment models
Constructing smart financial portfolios from data driven quantitative investment models
Portfolio managers have access to large amounts of financial time series data, which is rich in structure and information. Such structure, at varying time horizons and frequencies, exhibits different characteristics, such as momentum and mean reversion to mention two. The key challenge in building a smart portfolio is to first, identify and model the relevant data regimes operating at different time frames and then convert them into an investment model targeting each regime separately. Regimes in financial time series can change over a period of time, i.e. they are heterogeneous. This has implications for a model, as it may stop being profitable once the regime it is targeting has stopped or evolved into another one over a period of time. Changing regimes or those evolving into other regimes is one of the key reasons why we should have several independent models targeting relevant regimes at a particular point in time.

In this thesis we present a smart portfolio management approach that advances existing methods and one that beats the Sharpe ratio of other methods, including the efficient frontier. Our smart portfolio is a two-tier framework. In the first tier we build four quantitative investment models, with each model targeting a pattern at different time horizon. We build two market neutral models using the pairs methodology and the other two models use the momentum approach in the equity market. In the second tier we build a set of meta models that allocate capital to tier one, using Kelly Criterion, to build a meta portfolio of quantitative investment models. Our approach is smart at several levels. Firstly, we target patterns that occur in financial data at different time horizons and create high probability investment models. Hence we make better use of data. Secondly, we calculate the optimal bet size using Kelly at each time step to maximise returns. Finally we avoid making investments in loss making models and hence make smarter allocation of capital.
Mehra, Chetan Saran
dfd50ace-2d70-4846-a4a9-2b36dd352bfa
Mehra, Chetan Saran
dfd50ace-2d70-4846-a4a9-2b36dd352bfa
Gerding, Enrico
d9e92ee5-1a8c-4467-a689-8363e7743362

Mehra, Chetan Saran (2016) Constructing smart financial portfolios from data driven quantitative investment models. University of Southampton, Faculty of Physical Science and Engineering, Doctoral Thesis, 261pp.

Record type: Thesis (Doctoral)

Abstract

Portfolio managers have access to large amounts of financial time series data, which is rich in structure and information. Such structure, at varying time horizons and frequencies, exhibits different characteristics, such as momentum and mean reversion to mention two. The key challenge in building a smart portfolio is to first, identify and model the relevant data regimes operating at different time frames and then convert them into an investment model targeting each regime separately. Regimes in financial time series can change over a period of time, i.e. they are heterogeneous. This has implications for a model, as it may stop being profitable once the regime it is targeting has stopped or evolved into another one over a period of time. Changing regimes or those evolving into other regimes is one of the key reasons why we should have several independent models targeting relevant regimes at a particular point in time.

In this thesis we present a smart portfolio management approach that advances existing methods and one that beats the Sharpe ratio of other methods, including the efficient frontier. Our smart portfolio is a two-tier framework. In the first tier we build four quantitative investment models, with each model targeting a pattern at different time horizon. We build two market neutral models using the pairs methodology and the other two models use the momentum approach in the equity market. In the second tier we build a set of meta models that allocate capital to tier one, using Kelly Criterion, to build a meta portfolio of quantitative investment models. Our approach is smart at several levels. Firstly, we target patterns that occur in financial data at different time horizons and create high probability investment models. Hence we make better use of data. Secondly, we calculate the optimal bet size using Kelly at each time step to maximise returns. Finally we avoid making investments in loss making models and hence make smarter allocation of capital.

PDF
__userfiles.soton.ac.uk_Users_ojl1y15_mydesktop_final thesis.pdf - Other
Available under License University of Southampton Thesis Licence.
Download (2MB)

More information

Published date: November 2016
Organisations: University of Southampton, Agents, Interactions & Complexity

Identifiers

Local EPrints ID: 404673
URI: http://eprints.soton.ac.uk/id/eprint/404673
PURE UUID: f69431cd-59a2-48d2-abcc-c03584c8e394
ORCID for Enrico Gerding: ORCID iD orcid.org/0000-0001-7200-552X

Catalogue record

Date deposited: 30 Jan 2017 15:40
Last modified: 07 Jun 2019 00:34

Export record

Download statistics

Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.

View more statistics

Atom RSS 1.0 RSS 2.0

Contact ePrints Soton: eprints@soton.ac.uk

ePrints Soton supports OAI 2.0 with a base URL of http://eprints.soton.ac.uk/cgi/oai2

This repository has been built using EPrints software, developed at the University of Southampton, but available to everyone to use.

We use cookies to ensure that we give you the best experience on our website. If you continue without changing your settings, we will assume that you are happy to receive cookies on the University of Southampton website.

×