Search Results (5)

We devise a hierarchical decision-making architecture for portfolio optimization on multiple markets. At the highest level a Deep Reinforcement Learning (DRL) agent selects among a number of discrete actions, representing low-level agents. For the low-level agents, we use a set of Hierarchical Risk Parity (HRP) and Hierarchical Equal Risk Contribution (HERC) models with different hyperparameters, which all run in parallel, off-market (in a simulation). The information on which the DRL agent decides which of the low-level agents should act next is constituted by the stacking of the recent performances of all agents. Thus, the modelling resembles a statefull, non-stationary, multi-arm bandit, where the performance of the individual arms changes with time and is assumed to be dependent on the recent history. We perform experiments on the cryptocurrency market (117 assets), on the stock market (46 assets) and on the foreign exchange market (28 pairs) showing the excellent robustness and performance of the overall system. Moreover, we eliminate the need for retraining and are able to deal with large testing sets successfully. Full article

Poor feeding practices in infants and young children may lead to malnutrition, which, in turn, is associated with an increased risk of infectious diseases, such as respiratory tract infections (RTIs), a leading cause of under-five mortality. We explored the association between RTIs and the WHO infant and young child feeding (IYCF) indicators: minimum dietary diversity (MDD), minimum meal frequency (MMF), and minimum acceptable diet (MAD), among infants and preschool children in Suriname. A validated pediatric food frequency questionnaire was used and data on RTIs, defined as clinical care for fever with respiratory symptoms, bronchitis, or pneumonia were obtained. Associations between feeding indicators and RTIs were explored using hierarchical logistic regression. Of 763 children aged 10–33 months, 51.7% achieved the MDD, 88.5% the MMF, and 46.5% the MAD. Furthermore, 73% of all children experienced at least one upper and/or lower RTI. Children meeting the MDD and MAD had significantly lower odds on RTIs (OR 0.53; 95%CI: 0.37–0.74, p < 0.001; OR 0.55; 95%CI: 0.39–0.78, p < 0.001, respectively). The covariates parity and household income were independently associated with RTIs. In conclusion, MDD and MAD were associated with (upper) RTIs. Whether these indicators can be used as predictors for increased risk for RTIs should be assessed in future prospective studies. Full article

Background. France implemented in 2004 the French National Breast Cancer Screening Programme (FNBCSP). Despite national recommendations, this programme coexists with non-negligible opportunistic screening practices. Aim. Analyse socio-territorial inequities in the 2013–2014 FNBCSP campaign in a large sample of the eligible population. Method. Analyses were performed using three-level hierarchical generalized linear model. Level one was a 10% random sample of the eligible population in each département (n = 397,598). For each woman, age and travel time to the nearest accredited radiology centre were computed. These observations were nested within 22,250 residential areas called “Îlots Regroupés pour l’Information Statistique” (IRIS), for which the European Deprivation Index (EDI) is defined. IRIS were nested within 41 départements, for which opportunistic screening rates and gross domestic product based on purchasing power parity were available, deprivation and the number of radiology centres for 100,000 eligible women were computed. Results. Organized screening uptake increased with age (OR_1SD = 1.05 [1.04–1.06]) and decreased with travel time (OR_1SD = 0.94 [0.93–0.95]) and EDI (OR_1SD = 0.84 [0.83–0.85]). Between départements, organized screening uptake decreased with opportunistic screening rate (OR_1SD = 0.84 [0.79–0.87]) and départements deprivation (OR_1SD = 0.91 [0.88–0.96]). Association between EDI and organized screening uptake was weaker as opportunistic screening rates and as département deprivation increased. Heterogeneity in FNBCSP participation decreased between IRIS by 36% and between départements by 82%. Conclusion. FNBCSP does not erase socio-territorial inequities. The population the most at risk of dying from breast cancer is thus the less participating. More efforts are needed to improve equity. Full article

The Hierarchical risk parity (HRP) approach of portfolio allocation, introduced by Lopez de Prado (2016), applies graph theory and machine learning to build a diversified portfolio. Like the traditional risk-based allocation methods, HRP is also a function of the estimate of the covariance matrix, however, it does not require its invertibility. In this paper, we first study the impact of covariance misspecification on the performance of the different allocation methods. Next, we study under an appropriate covariance forecast model whether the machine learning based HRP outperforms the traditional risk-based portfolios. For our analysis, we use the test for superior predictive ability on out-of-sample portfolio performance, to determine whether the observed excess performance is significant or if it occurred by chance. We find that when the covariance estimates are crude, inverse volatility weighted portfolios are more robust, followed by the machine learning-based portfolios. Minimum variance and maximum diversification are most sensitive to covariance misspecification. HRP follows the middle ground; it is less sensitive to covariance misspecification when compared with minimum variance or maximum diversification portfolio, while it is not as robust as the inverse volatility weighed portfolio. We also study the impact of the different rebalancing horizon and how the portfolios compare against a market-capitalization weighted portfolio. Full article

Dystocia and perinatal mortality are major animal health, welfare and economic issues in beef suckler cow production. The objective of this study was to identify risk factors for dystocia and perinatal mortality and to analyze the relationships of both traits to external pelvic parameters in extensively kept beef suckler cows. Calving ease and calf survival were recorded for 785 births on five Angus cattle farms in Germany. The prevalence of dystocia and perinatal mortality was 3.4% and 4.3%, respectively. A hierarchical model was used to predict dystocia and perinatal mortality. First-parity dams had a higher probability of dystocia (p < 0.0001) than later-parity ones. Increasing birth weight was associated with an increasing risk for dystocia (p < 0.05). The probability of perinatal mortality (p < 0.0001) was higher in assisted births than in unassisted births. Calves from first-parity dams had a higher risk (p < 0.01) of being stillborn than calves from dams in later parities. An increase in the length of the pelvis was associated with an increase in odds for perinatal mortality (p < 0.001). In conclusion, the study indicates that dystocia and perinatal mortality are mainly problems in first-parity suckler cows. Concerning the predictive value of external pelvic parameters, further research is necessary. Full article