Bachelor's Degree Project Presentations: Lia Arvidsson, Martina Zanetti, Karolina Rakic, Svante Sirviö, Kristoffer Kulju and Edo Carlson

Lia Arvidsson

Time: 9:15-11:00

Place: 80115

Title: Comparing Long-Term Performance in Swedish Cross-country Skiers Through the Analysis of FIS-points

Abstract: In this project, we examine the performance development of Swedish cross-country skiers, specifically comparing athletes who were selected for the national team with those who were not. The study covers both sprint and distance skiing categories. To understand the athletes' performance improvements over time, we used historical data from FIS-point lists, collected for the period from 2000 to 2025, focusing on skiers between 17 and 27 years old. We applied Emax-models to estimate two main parameters for the different groups: the rate of development (E50) and the achieved peak performance (Emax). Interestingly, the statistical tests showed no significant difference in the development rate among the groups. Both national team and non-selected skiers reached half of their peak performance (E50) at approximately 20 years of age. However, there were clear, significant differences in the Emax parameter, showing that elite athletes ultimately achieve a higher final level, even though their growth curve follows the same timeline. Based on these results, the thesis concludes by calculating a one-sided performance threshold for 20-year-olds. This threshold is proposed as a practical tool for talent identification, helping coaches identify young athletes who have the potential to reach the national team level.

Martina Zanetti

Time: 10:15-11:00

Place: 80115

Title: Comparison of Models for Setting Analytical Performance Specifications in External Quality Assessment of Haematological Analyses

Abstract: This presentation addresses the comparison of different models used for setting analytical performance specifications (APS) in external quality assessment of haematological analyses. Traditionally, the total allowable error (TEa) model has been widely applied, combining both systematic and random error, referred to as bias and imprecision, respectively. However, the total error concept has increasingly been questioned, and alternative approaches based on measurement uncertainty, such as allowable expanded measurement uncertainty (MAU), have been proposed, placing greater emphasis on imprecision. The presentation will compare how these models classify laboratory results and further explore laboratory performance through estimation of bias and imprecision.

Karolina Rakic

Time: 11:15-12:00

Place: 80115

Title: Analysis of Offensive Possession Sequences in BK Häcken

Abstract: This thesis analyzes offensive possession sequences in BK Häcken using logistic regression. Using event data from the 2025 Allsvenskan season, a model was developed to estimate the probability that an offensive sequence results in a goal. The results show that progressive passes, dribbles, and crosses were strongly associated with offensive threat. The model was also used to identify dangerous possession sequences and analyze player involvement in offensive situations.

Svante Sirviö

Time: 13:15-14:00

Place: 80115

Title: Travel time tomography and the X-ray transform: A model case of boundary rigidity

Abstract: How can we learn the inner structure of an object without looking inside it? In tomography, we try to answer this question using different measurements on the boundary of the object. Travel time tomography measures the times it takes for particles or waves to travel between different points on the boundary of the object. We model the objects as smooth Manifolds, and the travel times as the distance induced by a Riemannian metric. In this thesis, we discuss why we cannot determine the metric exactly from only the boundary distance function. We study a Euclidean model of the problem and prove a simple boundary rigidity result using a one-sided conformal assumption. We then study what happens if we make small perturbations of the metric and see that this leads to the Radon transform and the X-ray transform.

Kristoffer Kulju

Time: 14:15-15:00

Place: 80115

Title: Fourier Analysis of ARMA series

Abstract: This thesis seeks to study ARMA series in the frequency domain. Using tools from the study of time-series as well as Fourier Analysis, we define the Power Spectral Density (PSD) and interpret it as variance accumulation across different frequencies. We define the periodogram as an estimator for the PSD, and through Herglotz's Theorem consider the more general notion of Power Spectral Measure. We conclude by studying when stationarity breaks down, leading us naturally to consider ARIMA and ARFIMA series.

Edo Carlson

Time: 15:15-16:00

Place: 80115

Title: Hypothesis Testing with Left-Censored Data: A Comparison of Naive and Censoring-Aware Methods for Two-Group and Multiple-Group Settings

Abstract: This thesis investigates hypothesis testing for left-censored data, where observations below a limit of detection are only partially observed. Through simulation studies for two-group and multiple-group comparisons, conventional substitution methods are compared with censoring-aware methods implemented in \texttt{R}. The results show that replacing censored observations with the LoD or LoD/2 can introduce systematic bias, leading to distorted p-values, inflated Type I error, and unreliable power. In contrast, censoring-aware methods generally provide more stable inference, especially under low to moderate censoring. The findings support the use of statistical methods specifically designed for censored data rather than simple substitution approaches.