Presentation Title

Algorithmic Portfolio Management

Faculty Mentor

Sarah Cannon

Start Date

23-11-2019 10:45 AM

End Date

23-11-2019 11:30 AM

Location

122

Session

poster 4

Type of Presentation

Poster

Subject Area

business_economics_public_administration

Abstract

Building investment portfolios for a variety of people with diverse needs is challenging. It can be the primary monetary source that sustains the livelihoods of myriads of people after retirement. Therefore, accuracy in picking investment strategies, that are likely to reach investor’s goals, is imperative. Furthermore, for investors with higher risk tolerance, building a profitable portfolio becomes an even greater challenge wherein unforeseen events can lead to major losses.

In this project, we aim to design an algorithm that creates an optimal investment portfolio based on an investor’s risk tolerance, time horizon, budget and target return. Safer securities would include bonds or index funds, while riskier investments entail stocks or derivatives. Our research aims to produce a portfolio that maximizes investment utility given a variety of constraints. We aim to design an algorithm capable of shortlisting various types of investments, in order to tailor each portfolio to fit various risk levels and budgets. The algorithm uses a capital asset pricing model to predict the returns of varying investment bundles and then choose the optimal range of investments to form the portfolio and return these securities as the output. This output will list investments purchased at differing levels depending on the investor’s given preferences. We hope to use our research to develop an algorithmic approach to portfolio management.

We foresee that the major challenges with the algorithm will be the accuracy in which it can predict the outcomes of different investment bundles as well as the lack of an optimal exit point for stock investment. However, given enough research and algorithm recursion we plan to minimize the variability of the predictions.

This document is currently not available here.

Share

COinS
 
Nov 23rd, 10:45 AM Nov 23rd, 11:30 AM

Algorithmic Portfolio Management

122

Building investment portfolios for a variety of people with diverse needs is challenging. It can be the primary monetary source that sustains the livelihoods of myriads of people after retirement. Therefore, accuracy in picking investment strategies, that are likely to reach investor’s goals, is imperative. Furthermore, for investors with higher risk tolerance, building a profitable portfolio becomes an even greater challenge wherein unforeseen events can lead to major losses.

In this project, we aim to design an algorithm that creates an optimal investment portfolio based on an investor’s risk tolerance, time horizon, budget and target return. Safer securities would include bonds or index funds, while riskier investments entail stocks or derivatives. Our research aims to produce a portfolio that maximizes investment utility given a variety of constraints. We aim to design an algorithm capable of shortlisting various types of investments, in order to tailor each portfolio to fit various risk levels and budgets. The algorithm uses a capital asset pricing model to predict the returns of varying investment bundles and then choose the optimal range of investments to form the portfolio and return these securities as the output. This output will list investments purchased at differing levels depending on the investor’s given preferences. We hope to use our research to develop an algorithmic approach to portfolio management.

We foresee that the major challenges with the algorithm will be the accuracy in which it can predict the outcomes of different investment bundles as well as the lack of an optimal exit point for stock investment. However, given enough research and algorithm recursion we plan to minimize the variability of the predictions.