Statistical testing of claims related to high-frequency stock market data

Abstract

Nowadays one can find a number of sources, presenting their views on possible stock price behavior patterns. The objective of this thesis is to interpret some claims regarding stock price fluctuations into mathematical formulations. The latter are further tested through relevant statistical tests in two stages. The first stage includes running the test on the dataset of one trading day per financial instrument. The second stage considers running a statistical test on the outcomes of the first stage in order to finally approve or reject the claim. Trading simulations will be run through the datasets of confirmed claims. The main value of the thesis is the demonstration of various possibilities to interpret the imprecise claims into mathematical formulations and then verify them by scientific approach.