Πρόβλεψη και ανάλυση της απόδοσης σε αγώνες αυτοκίνητου Formula 1
Prediction and analysis of performance in Formula 1 races
View/
Keywords
Data Analysis ; Prediction ; Timeseries ; Classification ; Clustering ; Machine Learning ; Regression ; Formula 1Abstract
Formula 1 also known as the pinnacle of motorsport, combines cutting-edge technology with world-class driver talent and strategic team tactics in a sport where milliseconds will make all the difference between victory and defeat. As races unfold unpredictably according to weather, tire strategies, on-track mishaps and fluctuating car performance, Formula 1 presents a fertile field for data analysis. Teams utilize advanced analytics to maximize their race strategy, to improve car performance and optimize pit stops. Beyond that, the growing availability of race data has enabled researchers and analysts to utilize advanced statistical methods to find trends, to analyze driver and team performance and to develop predictive models. With the use of machine learning and applied statistics, it is possible to move beyond traditional race understandings and uncover deeper patterns that define success in the sport.
The current thesis delves into the application of statistical and machine learning methods to study and predict Formula 1 racing performance. The research is preceded by extensive data preparation and feature creation, such as exploratory analysis via descriptive statistics, correlation analysis, and normality tests. Time series forecasting and regression techniques facilitate tracking and prediction of performance trends using historic data. Dimensionality reduction is achieved through implementation of Principal Component Analysis (PCA) in the machine learning chapter. Random Forest and Support Vector Machines (SVM) classification models are employed in the prediction of race outcomes. Clustering techniques such as K-Means and Hierarchical Clustering identify patterns in driver performance. The results demonstrate the effectiveness of these methodologies for determining key performance indicators and trends, outlining how statistical and machine learning methodologies used can enhance our understanding of motorsport analytics by providing data-driven insights into race performance and competitiveness.
Postgraduate Studies Programme
Εφαρμοσμένη ΣτατιστικήDepartment
Σχολή Χρηματοοικονομικής και Στατιστικής. Τμήμα Στατιστικής και Ασφαλιστικής ΕπιστήμηςNumber of pages
223Language
EnglishCollections
Except where otherwise noted, this item's license is described as
Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα
Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα