Search

The identification of movement strategies in situations that are as ecologically valid as possible is essential for the understanding of lower limb interactions. This study considered the kinetic and kinematic data for the hip, knee and ankle joints from 376 block jump-landings when moving in the dominant and non-dominant directions from fourteen senior national female volleyball players. Two Machine Learning methods were used to generate the models from the dataset, Random Forest and Artificial Neural Networks. In addition, decision trees were used to detect which variables were relevant to discern the limb movement strategies and to provide a meaningful prediction. The results showed statistically significant differences when comparing the movement strategies between limb role (accuracy > 88.0% and > 89.3%, respectively), and when moving in the different directions but performing the same role (accuracy > 92.3% and > 91.2%, respectively). This highlights the importance of considering limb dominance, limb role and direction of movement during block jump-landings in the identification of which biomechanical variables are the most influential in the movement strategies. Moreover, Machine Learning allows the exploration of how the joints of both limbs interact during sporting tasks, which could provide a greater understanding and identification of risky movements and preventative strategies. All these detailed and valuable descriptions could provide relevant information about how to improve the performance of the players and how to plan trainings in order to avoid an overload that could lead to risk of injury. This highlights that, there is a necessity to consider the learning models, in which the spike approach unilaterally is taught before the block approach (bilaterally). Therefore, we support the idea of teaching bilateral approach before learning the spike, in order to improve coordination and to avoid asymmetries between limbs. [ABSTRACT FROM AUTHOR]

Congenital dyserythropoietic anemias (CDA) are disorders characterized by ineffective erythropoiesis and morphological anomalies in erythrocytes and erythroblasts. The purpose of this study is to identify the gene variants in patients diagnosed with CDA. We analyzed five unrelated patients and two siblings with a targeted panel of genes to CDA: CDAN1, CDIN1, SEC23B, KIF23, KLF1, and GATA1 genes. We found three novel variants in the CDIN1 gene (p.Leu136Val, p.Tyr247Cys, and p.Ile273Thr), four known variants in the SEC23B gene (p.Arg14Trp, p.Arg554Ter, p.Asp239Gly, and p.Ser436Leu), and one novel variant in the KIF23 gene (p.Leu945Trpfs*31). The in silico analysis of novel variants predict that they are pathogenic and, the in vitro study confirms the functional impact of the KIF23 variant on the protein location. [ABSTRACT FROM AUTHOR]

An increasing amount of unlabeled time series data available render the semi-supervised paradigm a suitable approach to tackle classification problems with a reduced quantity of labeled data. Self-labeled techniques stand out from semi-supervised classification methods due to their simplicity and the lack of strong assumptions about the distribution of the labeled and unlabeled data. This paper addresses the relevance of these techniques in the time series classification context by means of an empirical study that compares successful self-labeled methods in conjunction with various learning schemes and dissimilarity measures. Our experiments involve 35 time series datasets with different ratios of labeled data, aiming to measure the transductive and inductive classification capabilities of the self-labeled methods studied. The results show that the nearest-neighbor rule is a robust choice for the base classifier. In addition, the amending and multi-classifier self-labeled-based approaches reveal a promising attempt to perform semi-supervised classification in the time series context. [ABSTRACT FROM AUTHOR]