Μελέτη οικονομικών στοιχείων για την πρόβλεψη συμπεριφοράς τους στη σύγχρονη οικονομική πραγματικότητα με χρήση τεχνητών νευρωνικών δικτύων

Ξενάκης, Κωνσταντίνος

Study of financial data to predict their behavior in modern economic reality using artificial neural networks

Master Thesis

Author

Ξενάκης, Κωνσταντίνος

Date

2024-02

Abstract

The economic benefits of accurately predicting price movements in the stock market are obvious. This task has traditionally been solved by analyzing the underlying companies and their historical share prices. A third option is to build a stock price forecasting model using machine learning. Following the latter approach, this paper presents a machine learning algorithm that uses historical stock price data. The algorithm uses this information to train a model that infer future stock prices based on recent stock price information. Machine learning is an important branch of computer science and is in constant development. Dramatic advances in the class of machine learning models known as artificial neural networks have led to increased interest in such models, including their application to financial forecasting. With so many models available, it is difficult to choose among them, especially given that new models and learning techniques are constantly emerging. Initially, the first chapter presents the concept of machine learning and its types. In the second chapter of the paper, the concept of the stock market is analyzed. The rules that govern it, the elements that are necessary to be known so that there is a full understanding of the creation of the forecasting models. Subsequently, the third chapter was devoted to the analysis of the definition of forecasting and the principles governing it. The most basic forecasting methods are presented and finally the evaluation measures and possible errors are examined. In the fourth chapter, the main time series are approached which, in combination with the selected predictive model, will constitute the algorithm to set up the neural network for predicting the price of a share. In the fifth chapter, a theoretical introduction to neural networks is given, their architecture is presented and the most basic neural artificial networks are analyzed. Finally, in the sixth chapter, the experimental control will be done with the use of appropriate software and, closing, the conclusive approach to the results will be made.

Postgraduate Studies Programme

Προηγμένα Συστήματα Πληροφορικής

Department

Σχολή Τεχνολογιών Πληροφορικής και Επικοινωνιών. Τμήμα Ψηφιακών Συστημάτων

Number of pages

Language

Greek

URI

https://dione.lib.unipi.gr/xmlui/handle/unipi/17258
http://dx.doi.org/10.26267/unipi_dione/4681

Collections

Τμήμα Ψηφιακών Συστημάτων

Show full item record

Except where otherwise noted, this item's license is described as
Αναφορά Δημιουργού-Όχι Παράγωγα Έργα 3.0 Ελλάδα