DSpace JSPUI
DSpace preserves and enables easy and open access to all types of digital content including text, images, moving images, mpegs and data sets
Learn More
Please use this identifier to cite or link to this item:
http://dspace.dtu.ac.in:8080/jspui/handle/repository/18001| Title: | MACHINE LEARNING FOR PORTFOLIO MANAGEMENT |
| Authors: | KS, LAKSHMI GUPTA, RASHIKA |
| Keywords: | Investment portfolio management portfolio Equity Shares |
| Issue Date: | 18-Aug-2020 |
| Abstract: | Artificial Intelligence & Machine Learning has already established a fairly strong foothold in the field of finance & other associated areas. On the front end machine learning is widely utilized for risk management and fraudulent transaction detection. At the front end AI is used for customer segmentation and support and pricing the derivatives (options ,futures & such). The only arena with limited Machine Learning usage has been in the buy-side of financial activity or more precisely Portfolio Management which includes selection of the best portfolio among a set of portfolios in accordance with certain objective such as expected return, financial risk in short any tangible or intangible aim. The aim of this project is to review traditional mathematical methods for portfolio optimization ( such as Markowitz ) , unsupervised ( such as Principal Component Analysis), supervised machine learning approaches, and related techniques. For the purpose of this project we will be using NSE stock data and identifying the top stock options and applying various optimization algorithms using python. Based on the results obtained we plan on identifying the optimal method. |
| Description: | Investment and portfolio management is a fairly significant aspect of the secondary capital market due to the fact that it helps in mobilizing the savings of the investors and subsequently assisting in the development of the economy via saving and transfer process. An investor can be a corporation , an individual, a government, or a pension fund. There are a variety of investment options available, offering different risk-reward tradeoffs. A thorough understanding of the core concepts and analysis of the various options available can help the investors in creating a portfolio the maximizes returns while reducing the exposure to risk. With the availability of a wide variety of investment aveneues , investors have considerable options to build their portfolio after weighing the pros and cons of each option. There are two main categories of investment options i.e., Financial assets : equity shares, derivative instruments , government securities, post office schemes, mutual fund shares, corporate debentures ,insurance policies, and deposit with banks. Real assets : tangible assets such as house, gold, commercial property, agricultural farm, precious stones, and art objects. There are two broad categories of investing – direct and indirect investing . Direct investing is where investors manage their own individual portfolios and the risk and return they receive is solely dependent on their ability to analyze the market behaviour and fluctuations. Indirect investing involves financial intermediaries which invest pools of funds into the market and maintain the investor portfolios, providing the investors with expert advice and recommendations and relieving the investors of making their own decisions. This is the point where portfolio managers come in, their primary job is to manage investor portfolios framed according to their invidual preferences ,return objectives and risk bearing capacity. Portfolio management entails portfolio planning, identification, selection and construction, feedback and evaluation of securities. The hidden talent in portfolio management lies in obtaining an adequate balance between the objectives of safety, liquidity and profitability. Traditionally after setting up the investment policy and portfolio objectives by assessing the current and future financial needs of the individual investor, the next task for the investor/portfolio manager is the analysis and evaluation of the investment options performed with the help of a combination of technical and fundamental analysis. Technical analysis aims to predict the future movement of the price of a particular financial asset that is traded on the market , one of the main drawback of technical analysis is the fact that the analysis of historical prices is based on the assumption that trends and patterns repeat itself. On the other hand fundamental analysis aims to determne the intrinsic value of an particular financial asset, it helps in determining which financial asset is over-priced or under- priced based on the difference between their market value and intrinsic value. Similar to technical analysis , fundamental analysis is also prone to errors and bias one of the assumptions that gives rise to the same is the fact that fundamental analysis assumes that intrinsic value is the present value of future flows from particular investment. Thus there was felt a need to provide sound and accurate analysis of investment instruments in order to construct even more risk - resistant and high return yielding portfolios. Modern portfolio theory (MPT) gave rise to the same. MPT or portfolio theory was first introduced by Harry Markowitz in his paper “Portfolio Selection” in the Journal of Finance ( 1952 ). Before modern portfolio theory ( MPT ), the decision about whether to include a investment option in a portfolio was based solely upon the fundamental analysis of the firm, its dividend policy and its financial statements. Harry Markowitz stirred a whirlwing by suggesting the fact that the value of a security to an investor might can be evaluated optimally by calculating its mean, its standard deviation and its correlation to other securities in the portfolio. Portfolio theory evaluates how risk - averse investors frame portfolios in order to optimize expected returns for a given level of market risk. The theory also aims to quantify the benefits of diversification. Portfolio theory constructs an efficient frontier of optimal portfolios out of a universe of risky assets . Every portfolio on the efficient frontier provides the maximum expected return for a particular level of risk. Investors are required to hold one of the optimal portfolios on the efficient frontier and adjust their total market risk positions within the risk - free financial asset. Furthermore in this study we have sought to explore the applications of Machine Learning algorithms in the field of investment and portfolio management Machine learning based methods that refer to statistical learning with data are widely applicable in computational finance.Which is found to be particularly helpful in order to obtain more accurate and risk resistant portfolios and also to overcome the limitations and shortcomings of the traditional portfolio optimization techiques. Some of the Machine Learning lagorithms that we have tried to use are Principal Component Analysis (PCA), Auto-encoder risk, Hierarchical Clustering, etc |
| URI: | http://dspace.dtu.ac.in:8080/jspui/handle/repository/18001 |
| Appears in Collections: | MBA |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| machine learning for portfolio management.pdf | Investment and portfolio management is a fairly significant aspect of the secondary capital market due to the fact that it helps in mobilizing the savings of the investors and subsequently assisting in the development of the economy via saving and transfer process. An investor can be a corporation , an individual, a government, or a pension fund. There are a variety of investment options available, offering different risk-reward tradeoffs. A thorough understanding of the core concepts and analysis of the various options available can help the investors in creating a portfolio the maximizes returns while reducing the exposure to risk. With the availability of a wide variety of investment aveneues , investors have considerable options to build their portfolio after weighing the pros and cons of each option. There are two main categories of investment options i.e., Financial assets : equity shares, derivative instruments , government securities, post office schemes, mutual fund shares, corporate debentures ,insurance policies, and deposit with banks. Real assets : tangible assets such as house, gold, commercial property, agricultural farm, precious stones, and art objects. There are two broad categories of investing – direct and indirect investing . Direct investing is where investors manage their own individual portfolios and the risk and return they receive is solely dependent on their ability to analyze the market behaviour and fluctuations. Indirect investing involves financial intermediaries which invest pools of funds into the market and maintain the investor portfolios, providing the investors with expert advice and recommendations and relieving the investors of making their own decisions. This is the point where portfolio managers come in, their primary job is to manage investor portfolios framed according to their invidual preferences ,return objectives and risk bearing capacity. Portfolio management entails portfolio planning, identification, selection and construction, feedback and evaluation of securities. The hidden talent in portfolio management lies in obtaining an adequate balance between the objectives of safety, liquidity and profitability. Traditionally after setting up the investment policy and portfolio objectives by assessing the current and future financial needs of the individual investor, the next task for the investor/portfolio manager is the analysis and evaluation of the investment options performed with the help of a combination of technical and fundamental analysis. Technical analysis aims to predict the future movement of the price of a particular financial asset that is traded on the market , one of the main drawback of technical analysis is the fact that the analysis of historical prices is based on the assumption that trends and patterns repeat itself. On the other hand fundamental analysis aims to determne the intrinsic value of an particular financial asset, it helps in determining which financial asset is over-priced or under- priced based on the difference between their market value and intrinsic value. Similar to technical analysis , fundamental analysis is also prone to errors and bias one of the assumptions that gives rise to the same is the fact that fundamental analysis assumes that intrinsic value is the present value of future flows from particular investment. Thus there was felt a need to provide sound and accurate analysis of investment instruments in order to construct even more risk - resistant and high return yielding portfolios. Modern portfolio theory (MPT) gave rise to the same. MPT or portfolio theory was first introduced by Harry Markowitz in his paper “Portfolio Selection” in the Journal of Finance ( 1952 ). Before modern portfolio theory ( MPT ), the decision about whether to include a investment option in a portfolio was based solely upon the fundamental analysis of the firm, its dividend policy and its financial statements. Harry Markowitz stirred a whirlwing by suggesting the fact that the value of a security to an investor might can be evaluated optimally by calculating its mean, its standard deviation and its correlation to other securities in the portfolio. Portfolio theory evaluates how risk - averse investors frame portfolios in order to optimize expected returns for a given level of market risk. The theory also aims to quantify the benefits of diversification. Portfolio theory constructs an efficient frontier of optimal portfolios out of a universe of risky assets . Every portfolio on the efficient frontier provides the maximum expected return for a particular level of risk. Investors are required to hold one of the optimal portfolios on the efficient frontier and adjust their total market risk positions within the risk - free financial asset. Furthermore in this study we have sought to explore the applications of Machine Learning algorithms in the field of investment and portfolio management Machine learning based methods that refer to statistical learning with data are widely applicable in computational finance.Which is found to be particularly helpful in order to obtain more accurate and risk resistant portfolios and also to overcome the limitations and shortcomings of the traditional portfolio optimization techiques. Some of the Machine Learning lagorithms that we have tried to use are Principal Component Analysis (PCA), Auto-encoder risk, Hierarchical Clustering, etc | 2.75 MB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.