Back to search

National and Kapodistrian University of Athens

Methodological advancements in time series analysis & forecasting

Abstract

dc:description

This thesis comprises four contributions, each presenting a novel methodology for financial time series analysis and forecasting. The research builds on the concept of maximum entropy and its application to portfolio analysis, time series forecasting, and smoothing, as well as quantitative trading strategies. Specifically, in the first contribution (Chapter 2), we modify the weights of the standard portfolio benchmark, known as the equally weighted portfolio, under an adaptive learning method (based on a single parameter) to overcome the “curse of dimensionality” and achieve superior economic performance as evaluated by established trading metrics. In the following contribution (Chapter 3), we present a forecasting approach to market timing and examine how the most parsimonious models —i.e., autoregressive and constant models —can lead us to the concept of predictable profitability. In the final two contributions, in Chapters 4 and 5, we develop a new methodology for entropy ranking portfolios and maximum entropy forecasting. In particular, in Chapter 4, we associate the concept of entropy with the extraction of suitable probabilities to construct portfolio weights, having an explicit emphasis on a model-free, data-driven approach with minimal assumptions on the underlying data-generating process of financial returns. Then, in Chapter 5, we determine the weights of the original θ-forecast function. This Chapter provides a useful generalization of the well-known forecasting benchmark, the θ-method, based on an alternative approach to the θ coefficient through the principle of maximum entropy, thereby expanding its scope and applicability. We conclude by presenting, in the appendix, the application of maximum entropy in the context of smoothing stationary time series.

Degree

thesis:*
Grantor dc:publisher
National and Kapodistrian University of Athens
Year dc:date
2025

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Tarani, Sophia
  • Ταράνη, Σοφία

Subjects

dc:subject × 44

Rights

Language dc:language
eng

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/58795

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Tarani, Sophia; Ταράνη, Σοφία. Methodological advancements in time series analysis & forecasting. National and Kapodistrian University of Athens, 2025. http://hdl.handle.net/10442/hedi/58795