Insufficient physical activity presents a significant hazard to overall health, with sedentary lifestyles linked to a variety of health issues. Monitoring physical activity levels allows the recognition of patterns of sedentary behavior and the provision of coaching to meet the recommended physical activity standards. In this paper, we aim to address the problem of reducing the time consuming process of fitting classifiers when generating personalized models for a coaching application. The proposed approach consists of evaluating the effects of clustering participants based on their walking patterns and then recommending a unique model for each group. Each model consists of a random forest classifier with a different number of estimators each. The resulting approach reduces the fitting time considerably while keeping nearly the same classification performance as personalized models.
DOCUMENT
We present a novel architecture for an AI system that allows a priori knowledge to combine with deep learning. In traditional neural networks, all available data is pooled at the input layer. Our alternative neural network is constructed so that partial representations (invariants) are learned in the intermediate layers, which can then be combined with a priori knowledge or with other predictive analyses of the same data. This leads to smaller training datasets due to more efficient learning. In addition, because this architecture allows inclusion of a priori knowledge and interpretable predictive models, the interpretability of the entire system increases while the data can still be used in a black box neural network. Our system makes use of networks of neurons rather than single neurons to enable the representation of approximations (invariants) of the output.
LINK
BACKGROUND & AIMS: Sufficient protein intake is of great importance in hemodialysis (HD) patients, especially for maintaining muscle mass. Daily protein needs are generally estimated using bodyweight (BW), in which individual differences in body composition are not accounted for. As body protein mass is best represented by fat free mass (FFM), there is a rationale to apply FFM instead of BW. The agreement between both estimations is unclear. Therefore, the aim of this study is to compare protein needs based on either FFM or BW in HD patients.METHODS: Protein needs were estimated in 115 HD patients by three different equations; FFM, BW and BW adjusted for low or high BMI. FFM was measured by multi-frequency bioelectrical impedance spectroscopy and considered the reference method. Estimations of FFM x 1.5 g/kg and FFM x 1.9 g/kg were compared with (adjusted)BW x 1.2 and x 1.5, respectively. Differences were assessed with repeated measures ANOVA and Bland-Altman plots.RESULTS: Mean protein needs estimated by (adjusted)BW were higher compared to those based on FFM, across all BMI categories (P < 0.01) and most explicitly in obese patients. In females with BMI >30, protein needs were 69 ± 17.4 g/day higher based on BW and 45 ± 9.3 g/day higher based on BMI adjusted BW, compared to FFM. In males with BMI >30, protein needs were 51 ± 20.4 g/day and 23 ± 20.9 g/day higher compared to FFM, respectively.CONCLUSIONS: Our data show large differences and possible overestimations of protein needs when comparing BW to FFM. We emphasize the importance of more research and discussion on this topic.
DOCUMENT
