Bracketing Uncertainty in Clustering Under the Manifold Hypothesis
本文针对聚类中的不确定性问题，提出基于流形假设的聚簇方法（MBC），通过几何与样本量度结合定义不确定性区间，量化数据固有的模糊性。
0 citationsRead paper
Selected work
Representative Papers
Learning-Augmented Optimization for Strategic Two-Echelon Spare Parts Network Design
研究通过结合图神经网络、变邻域搜索和集合划分重组的方法，优化了两层备件网络设计，解决了高成本评估问题，并提高了节约率。
0 citationsRead paper
How exceptional was the Big Three era? Extremes and persistence in men's professional tennis
本文通过贝叶斯动态Bradley-Terry状态空间模型评估2003至2023年间男子网球三巨头时代的独特性，发现其独特在于三人长时间保持顶尖地位而非个人实力。
0 citationsRead paper
Learning Latent Memory States from Longitudinal Athlete Monitoring Data
This study addresses the challenge of modeling and reusing recent historical information in longitudinal athlete monitoring by introducing the “latent memory table” as a novel analytical unit. The approach employs a Transformer-based memory operator to map masked temporal windows into finite-dimensional states, which are aggregated into a statistical table amenable to storage, querying, and reuse. This framework unifies and generalizes classical techniques such as moving averages and principal component analysis, while emphasizing six key quality attributes—including restorability, personalization, and temporal consistency—and incorporates uncertainty quantification, Procrustes-based row reliability ensembles, and a composite quality index Q for evaluation. In the SoccerMon case study, the latent memory table achieves a quality index of 0.73, substantially outperforming conventional methods (approximately 0.40) and demonstrating incremental predictive value for certain health indicators.
0 citationsRead paper
Modelling Athletic Ageing Relative to an Estimated Performance Envelope
This study addresses the challenge of modeling sparse and irregular longitudinal performance data in athletes, where existing approaches often focus on population-average aging trends while neglecting individual deviations from their latent peak performance—termed the performance envelope. The authors propose RACE, a two-stage framework that first estimates an age-conditioned high-performance envelope across the population and then employs STAR (Shape Translation And Rotation), a nonlinear mixed-effects model, to represent individual trajectories as low-dimensional geometric transformations of this envelope—capturing shifts in level, timing, and pace. A key innovation lies in jointly modeling the performance envelope and individual aging parameters, embedding the correlation between level and pace within the random-effects covariance structure, thereby revealing how the envelope’s geometry governs parameter identifiability. Applied to MLB Statcast data, the method successfully disentangles ability level from aging pace, explaining why burst initiation rate identifies timing shifts whereas sprint speed does not.
0 citationsRead paper
Recent publications
Latest Papers
Bracketing Uncertainty in Clustering Under the Manifold Hypothesis
本文针对聚类中的不确定性问题，提出基于流形假设的聚簇方法（MBC），通过几何与样本量度结合定义不确定性区间，量化数据固有的模糊性。
0 citationsRead paper
Learning-Augmented Optimization for Strategic Two-Echelon Spare Parts Network Design
研究通过结合图神经网络、变邻域搜索和集合划分重组的方法，优化了两层备件网络设计，解决了高成本评估问题，并提高了节约率。
0 citationsRead paper
How exceptional was the Big Three era? Extremes and persistence in men's professional tennis
本文通过贝叶斯动态Bradley-Terry状态空间模型评估2003至2023年间男子网球三巨头时代的独特性，发现其独特在于三人长时间保持顶尖地位而非个人实力。
0 citationsRead paper
Learning Latent Memory States from Longitudinal Athlete Monitoring Data
This study addresses the challenge of modeling and reusing recent historical information in longitudinal athlete monitoring by introducing the “latent memory table” as a novel analytical unit. The approach employs a Transformer-based memory operator to map masked temporal windows into finite-dimensional states, which are aggregated into a statistical table amenable to storage, querying, and reuse. This framework unifies and generalizes classical techniques such as moving averages and principal component analysis, while emphasizing six key quality attributes—including restorability, personalization, and temporal consistency—and incorporates uncertainty quantification, Procrustes-based row reliability ensembles, and a composite quality index Q for evaluation. In the SoccerMon case study, the latent memory table achieves a quality index of 0.73, substantially outperforming conventional methods (approximately 0.40) and demonstrating incremental predictive value for certain health indicators.
0 citationsRead paper
Modelling Athletic Ageing Relative to an Estimated Performance Envelope
This study addresses the challenge of modeling sparse and irregular longitudinal performance data in athletes, where existing approaches often focus on population-average aging trends while neglecting individual deviations from their latent peak performance—termed the performance envelope. The authors propose RACE, a two-stage framework that first estimates an age-conditioned high-performance envelope across the population and then employs STAR (Shape Translation And Rotation), a nonlinear mixed-effects model, to represent individual trajectories as low-dimensional geometric transformations of this envelope—capturing shifts in level, timing, and pace. A key innovation lies in jointly modeling the performance envelope and individual aging parameters, embedding the correlation between level and pace within the random-effects covariance structure, thereby revealing how the envelope’s geometry governs parameter identifiability. Applied to MLB Statcast data, the method successfully disentangles ability level from aging pace, explaining why burst initiation rate identifies timing shifts whereas sprint speed does not.
0 citationsRead paper