A data-driven approach for modeling the behavior of stock prices

TL;DR

This paper introduces two data-driven methods for modeling stock prices, using probability distributions and stochastic processes, validated with historical Apple stock data and information theory metrics.

Contribution

It presents a novel combination of probabilistic and stochastic modeling approaches for stock price behavior analysis.

Findings

Both models effectively simulate stock price movements.

The models outperform baseline methods in accuracy.

Insights into stock price dynamics are gained through information theory metrics.

Abstract

In this paper, we describe two approaches to model the behavior of stock prices. The first approach considers the underlying probability distribution of day-to-day price differences. The second approach models the movement of the price as a stochastic birth-death process. We demonstrated the two approaches using historical opening prices of Apple inc. and compared the simulated prices from the two approaches to the actual ones using information theory metrics.