Training load responses modelling and model generalisation in elite sports

Abstract

This study aims to provide a transferable methodology in the context of sport performance modelling, with a special focus to the generalisation of models. Data were collected from seven elite Short track speed skaters over a three months training period. In order to account for training load accumulation over sessions, cumulative responses to training were modelled by impulse, serial and bi-exponential responses functions. The variable dose-response (DR) model was compared to elastic net (ENET), principal component regression (PCR) and random forest (RF) models, while using cross-validation within a time-series framework. ENET, PCR and RF models were fitted either individually ($$M_{I}$$ M I ) or on the whole group of athletes ($$M_{G}$$ M G ). Root mean square error criterion was used to assess performances of models. ENET and PCR models provided a significant greater generalisation ability than the DR model ($$p = 0.018$$ p = 0.018 , $$p < 0.001$$ p < 0.001 , $$p = 0.004$$ p = 0.004 and $$p < 0.001$$ p < 0.001 for $$ENET_{I}$$ E N E T I , $$ENET_{G}$$ E N E T G , $$PCR_{I}$$ P C R I and $$PCR_{G}$$ P C R G , respectively). Only $$ENET_{G}$$ E N E T G and $$RF_{G}$$ R F G were significantly more accurate in prediction than DR ($$p < 0.001$$ p < 0.001 and $$p < 0.012$$ p < 0.012 ). In conclusion, ENET achieved greater generalisation and predictive accuracy performances. Thus, building and evaluating models within a generalisation enhancing procedure is a prerequisite for any predictive modelling.

Using Global Navigation Satellite Systems for modeling Athletic Performances in elite football players

“…In a small sample context, using a larger population may lead to more robust estimates of parameter coefficients. One possible way could be to build models over a group of players instead of a model per player 32 . Our results did not confirm such benefits since there is no benefit to using player-specific models for predicting A-V profiles with the current data.…”

“…Using a ridge regularisation was motivated by the high dimensional context that may lead to unsteady multivariate linear models, excessively sensitive to an expanded space of solutions. Accordingly, ridge regularisation reduces the space of solutions while solving collinearity problems, which remains common in sports [30][31][32] . It thus prevents biased estimates through penalising estimates of correlated features 33,34 .…”

Factor Analysis and Regression Prediction Model of Swimmers’ Performance Structure Based on Mixed Genetic Neural Network

“…Then, due to the influence of the Second World War, the performance of 10 km freestyle had obvious fluctuations. 50s post-war new order was established and the world re-entered the stage of peace, the level of the steeplechase was developed by leaps and bounds, and the world record was constantly set, and now, it has exceeded the 8 minutes' mark [ 9 ]. Looking at the trend and future development of 10 km freestyle in the Olympic Games, there is still some room for the development of this sport [ 10 ].…”

Construction of Enterprise English Adaptive Learning Platform Based on Big Data Analysis

“…e way of educating and guiding students to absorb knowledge has also undergone a fundamental change with the rapid development of globalization and information technology [3]. However, the core educational objective of "optimizing students' learning styles and improving the rate of knowledge absorption" remains the same [4]. Recently, the Decision of the Central Committee of the Communist Party of China on Several Major Issues of Comprehensively Deepening Reform clearly proposed to accelerate the construction of an e ective linkage mechanism that actively uses modern information technology to gradually expand the information coverage of high-quality higher education resources [5].…”