Αξιολόγηση της αποτελεσματικότητας εργαλείων τεχνητής νοημοσύνης στην ανάλυση και πρόβλεψη δεδομένων

Τάσσου, Κωνσταντίνα

Master Thesis

Author

Τάσσου, Κωνσταντίνα

Date

2026

Abstract

The objective of this master’s thesis is to investigate the effectiveness of Artificial Intelligence (AI) tools in the field of data analysis and forecasting, as compared to traditional empirical analysis. For the purposes of the experimental process, the "Hotel Booking Demand Dataset"—comprising authentic hotel booking records—was utilized, with the primary goal of modeling and predicting the Average Daily Rate (ADR) variable. The research approach was structured into distinct methodological stages. Initially, thorough data pre-processing and cleaning were performed, involving the identification and management of outliers, as well as the appropriate transformation of variables to ensure the reliability of the results. Within the framework of statistical analysis, simple and multivariate regression models were applied to interpret the determinants that shape pricing policies in the hotel industry. Particular emphasis was placed on data segmentation, a strategy that contributed decisively to optimizing both the explanatory power and the predictive capacity of the generated models. In the subsequent stage, the analysis was extended through the use of advanced AI tools, specifically the ChatGPT and Gemini models. The comparative evaluation of these two approaches revealed fundamental differences in their problem-solving philosophies: while ChatGPT favored the use of linear models with a high degree of interpretability, Gemini adopted more complex non-linear techniques, achieving superior performance in terms of predictive accuracy. In conclusion, the findings of this thesis demonstrate that the choice of methodology depends on the objective of the analysis. Traditional statistical models offer high interpretability and a deeper understanding of the underlying data, whereas Artificial Intelligence models achieve greater predictive precision. Consequently, a hybrid combination of the two approaches is proposed as an effective strategy for decision-making environments.

Postgraduate Studies Programme

Βιομηχανική Διοίκηση και Τεχνολογία

Department

Σχολή Ναυτιλίας και Βιομηχανίας. Tμήμα Βιομηχανικής Διοίκησης και Tεχνολογίας

Number of pages

Language

Greek

URI

https://dione.lib.unipi.gr/xmlui/handle/unipi/19631

Collections

Τμήμα Βιομηχανικής Διοίκησης και Τεχνολογίας

Show full item record

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