Blood Test-Based Age Acceleration Is Inversely Associated with High-Volume Sports Activity.

PURPOSE: We develop blood test-based aging clocks and examine how these clocks reflect high-volume sports activity. METHODS: We use blood tests and body metrics data of 421 Hungarian athletes and 283 age-matched controls (mean age, 24.1 and 23.9 yr, respectively), the latter selected from a group of healthy Caucasians of the National Health and Nutrition Examination Survey (NHANES) to represent the general population ( n = 11,412). We train two age prediction models (i.e., aging clocks) using the NHANES dataset: the first model relies on blood test parameters only, whereas the second one additionally incorporates body measurements and sex. RESULTS: We find lower age acceleration among athletes compared with the age-matched controls with a median value of -1.7 and 1.4 yr, P < 0.0001. BMI is positively associated with age acceleration among the age-matched controls ( r = 0.17, P < 0.01) and the unrestricted NHANES population ( r = 0.11, P < 0.001). We find no association between BMI and age acceleration within the athlete dataset. Instead, age acceleration is positively associated with body fat percentage ( r = 0.21, P < 0.05) and negatively associated with skeletal muscle mass (Pearson r = -0.18, P < 0.05) among athletes. The most important blood test features in age predictions were serum ferritin, mean cell volume, blood urea nitrogen, and albumin levels. CONCLUSIONS: We develop and apply blood test-based aging clocks to adult athletes and healthy controls. The data suggest that high-volume sports activity is associated with slowed biological aging. Here, we propose an alternative, promising application of routine blood tests.

Session Recommendation via Recurrent Neural Networks over Fisher Embedding Vectors.

Recommendation services bear great importance in e-commerce, shopping, tourism, and social media, as they aid the user in navigating through the items that are most relevant to their needs. In order to build recommender systems, organizations log the item consumption in their user sessions by using different sensors. For instance, Web sites use Web data loggers, museums and shopping centers rely on user in-door positioning systems to register user movement, and Location-Based Social Networks use Global Positioning System for out-door user tracking. Most organizations do not have a detailed history of previous activities or purchases by the user. Hence, in most cases recommenders propose items that are similar to the most recent ones viewed in the current user session. The corresponding task is called session based, and when only the last item is considered, it is referred to as item-to-item recommendation. A natural way of building next-item recommendations relies on item-to-item similarities and item-to-item transitions in the form of "people who viewed this, also viewed" lists. Such methods, however, depend on local information for the given item pairs, which can result in unstable results for items with short transaction history, especially in connection with the cold-start items that recently appeared and had no time yet to accumulate a sufficient number of transactions. In this paper, we give new algorithms by defining a global probabilistic similarity model of all the items based on Random Fields. We give a generative model for the item interactions based on arbitrary distance measures over the items, including explicit, implicit ratings and external metadata to estimate and predict item-to-item transition probabilities. We exploit our new model in two different item similarity algorithms, as well as a feature representation in a recurrent neural network based recommender. Our experiments on various publicly available data sets show that our new model outperforms simple similarity baseline methods and combines well with recent item-to-item and deep learning recommenders under several different performance metrics.