UCLA

The stock market is a risky and popular attraction for those wishing to add a little extra to their income. Various models have been developed to predict stock movements; they range from complexity to the variables taken into account. Some models might also present performance instability depending on the industries they are simulated in. It then becomes difficult for investors to select and learn a profitable model in a short amount of time.

In this thesis, we develop simple models that aim at maximizing profit for investors in the gasoline industry. Our models focus on correlations between the companies as the bigger players in the industry are often influential enough to affect the stock prices of the smaller companies. Five chosen companies are paired up and several tests were done to determine the possible occurrence of certain patterns. Three evaluation methods are used to validate and finally, confirm the existence of those patterns.