Abstract

The number of casual investors allocating funds into financial exchanges has surged due to the increased availability of trading accounts on multiple platforms. These investors quite often invest only in one type of asset, stocks. Stocks are known to experience sudden market shifts and extreme volatility based on factors that the investors may not be able to control. Existing applications of data mining stocks perform well, but if the entire stock market performs poorly, investors can face severe losses. This study utilized a data mining tool that evaluates two other classes of investments: the commodities market and the currency exchange market. Three avenues of data mining were implemented as solutions, a neural network, logistic regression and a decision tree, to classify the buying and selling of investments. The results presented that unless in a bullish market scenario, utilizing a multi-asset portfolio with backed by a data mining tool can prove beneficial to an investor. In a bearish market, this study outlined how the performance of the multi-asset portfolio is drastically better than investing using a standalone stock classifier or investing in an index tracked product. In a volatile market, results showed that a multi-asset portfolio is competitive with a standalone stock classifier and in many scenarios even out performed. Overall, the data and resulting analysis provides a good basis for further research.

Publication Date

6-2017

Document Type

Thesis

Student Type

Graduate

Degree Name

Computer Science (MS)

Department, Program, or Center

Computer Science (GCCIS)

Advisor

Rajendra K Raj

Advisor/Committee Member

Leonid Reznik

Advisor/Committee Member

Carol Romanowski

Campus

RIT – Main Campus

Share

COinS