Maximal autocorrelation functions in functional data analysis

This paper proposes a new factor rotation for the context of functional principal components analysis. This rotation seeks to re-express a functional subspace in terms of directions of decreasing smoothness as represented by a generalized smoothing metric. The rotation can be implemented simply and we show on two examples that this rotation can improve the interpretability of the leading components.     

“My nest is full:” Intergenerational relationships at midlife

Incorporating a life course perspective, this qualitative study used focus groups to explore the experiences of midlife adults who were simultaneously providing support to emerging adult children and aging parents. Results indicated that adults situated in middle generations held beliefs that endorsed family-based responsibility to both younger and older members. Parents gladly supported children despite their longer transition to adulthood. Often unanticipated but accepted, provisions of care to aging parents were experienced with ambivalence — a joy and a burden. The transition of their parents to greater dependence helped participants gain insights into the terrain of late life and encouraged reflections about the intersection of aging, independence, and family responsibility. Participants expressed intentions to preserve their own independence and spare their children of caregiving burdens through self-directed actions. Implications focused on negotiations of family relationships around issues of independence and family responsibilities as a way to reduce intergenerational ambivalence.     

Aldrin and Dieldrin: A Reevaluation of the Cancer and Noncancer Dose‐Response Assessments

The dose‐response analyses of cancer and noncancer health effects of aldrin and dieldrin were evaluated using current methodology, including benchmark dose analysis and the current U.S. Environmental Protection Agency (U.S. EPA) guidance on body weight scaling and uncertainty factors. A literature review was performed to determine the most appropriate adverse effect endpoints. Using current methodology and information, the estimated reference dose values were 0.0001 and 0.00008 mg/kg‐day for aldrin and dieldrin, respectively. The estimated cancer slope factors for aldrin and dieldrin were 3.4 and 7.0 (mg/kg‐day)^?1, respectively (i.e., about 5‐ and 2.3‐fold lower risk than the 1987 U.S. EPA assessments). Because aldrin and dieldrin are no longer used as pesticides in the United States, they are presumed to be a low priority for additional review by the U.S. EPA. However, because they are persistent and still detected in environmental samples, quantitative risk assessments based on the best available methods are required. Recent epidemiologic studies do not demonstrate a causal association between aldrin and dieldrin and human cancer risk. The proposed reevaluations suggest that these two compounds pose a lower human health risk than currently reported by the U.S. EPA.     

Bootstrap bias corrections for ensemble methods

This paper examines the use of a residual bootstrap for bias correction in machine learning regression methods. Accounting for bias is an important obstacle in recent efforts to develop statistical inference for machine learning. We demonstrate empirically that the proposed bootstrap bias correction can lead to substantial improvements in both bias and predictive accuracy. In the context of ensembles of trees, we show that this correction can be approximated at only double the cost of training the original ensemble. Our method is shown to improve test set accuracy over random forests by up to 70% on example problems from the UCI repository.     

Local quadratic estimation of the curvature in a functional single index model

The nonlinear responses of species to environmental variability can play an important role in the maintenance of ecological diversity. Nonetheless, many models use parametric nonlinear terms which pre-determine the ecological conclusions. Motivated by this concern, we study the estimate of the second derivative (curvature) of the link function in a functional single index model. Since the coefficient function and the link function are both unknown, the estimate is expressed as a nested optimization. We first estimate the coefficient function by minimizing squared error where the link function is estimated with a Nadaraya-Watson estimator for each candidate coefficient function. The first and second derivatives of the link function are then estimated via local-quadratic regression using the estimated coefficient function. In this paper, we derive a convergence rate for the curvature of the nonlinear response. In addition, we prove that the argument of the linear predictor can be estimated root-n consistently. However, practical implementation of the method requires solving a nonlinear optimization problem, and our results show that the estimates of the link function and the coefficient function are quite sensitive to the choices of starting values.     

Perceptions of Communication and Monitoring Technologies Among Older Rural Women: Problem or Panacea?

The purpose of this study was to explore older rural women's perceptions of gerontechnologies, such as in-home monitoring and communication, which may be particularly useful in rural areas where support services are scarce. Forty-three women, ages 60–85 years, participated in focus group discussions. Qualitative analysis uncovered six primary themes: access and reliability, cost, privacy, usability, “use it or lose it” mentality, and caregiver burden. Results suggest that older women in rural areas are open to the potential of gerontechnologies to extend their ability to age in place but are skeptical about their actual implementation.