Browsing by Subject "nonparametric regression"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    The Complex Relationship Between Personality and Functioning
    (2020-07) Wright, Zara
    In recent decades, psychopathology research has established significant evidence in support of a dimensional diagnostic model, in which maladaptive personality traits underlie and predict clusters of mental health symptoms. In this framework, psychopathology may be defined as maladaptively high or low levels of a personality trait causing distress and/or impairment. This literature, however, has yet to characterize the specific relationship between these traits and impairments in functioning (e.g., physical functioning, social functioning, mental health functioning). The current study aims to address this gap in the literature by (a) augmenting the measurement of personality traits along their full range by integrating cognate traits from the “maladaptive” and “normative” personality literature onto unidimensional personality spectra; (b) modeling the nonparametric relationship between newfound personality traits with functioning; (c) explore how these relationships are moderated by age and sex; and (d) validating initial findings using replication and confirmatory procedures in a second sample. Data for this study were collected, using item-sampling techniques, from an online personality questionnaire where individuals self-selected to participate in exchange for feedback on their personality profiles. The overall sample included 214,420 people (split into two samples of 107,210 individuals each) from 223 countries. Results provide support for the replicability of the relationships between personality and functioning. Evidence suggests these relationships are not linear and monotonic, but rather optimal functioning occurs between the extreme ends of the trait. Age and/or sex play different roles in moderating these relationships depending on the personality trait of interest. Future research is needed to address measurement problems which interfere with measuring the full spectrum of each personality trait.
  • Thumbnail Image
    Item
    Contextual Bandits With Delayed Feedback Using Randomized Allocation
    (2020-05) Arya, Sakshi
    Contextual bandit problems are important for sequential learning in various practical settings that require balancing the exploration-exploitation trade-off to maximize total rewards. Motivated by applications in health care, we consider a multi-armed bandit setting with covariates and allow for delay in observing the rewards (treatment outcomes) as would most likely be the case in a medical setting. We focus on developing randomized allocation strategies that incorporate delayed rewards using nonparametric regression methods for estimating the mean reward functions. Although there has been substantial work on handling delays in standard multi-armed bandit problems, the field of contextual bandits with delayed feedback, especially with nonparametric estimation tools, remains largely unexplored. In the first part of the dissertation, we study a simple randomized allocation strategy incorporating delayed feedback, and establish strong consistency. Our setup is widely applicable as we allow for delays to be random and unbounded with mild assumptions, an important setting that is usually not considered in previous works. We study how different hyperparameters controlling the amount of exploration and exploitation in a randomized allocation strategy should be updated based on the extent of delays and underlying complexities of the problem, in order to enhance the overall performance of the strategy. We provide theoretical guarantees of the proposed methodology by establishing asymptotic strong consistency and finite-time regret bounds. We also conduct simulations and real data evaluations to illustrate the performance of the proposed strategies. In addition, we consider the problem of integrating expert opinion into a randomized allocation strategy for contextual bandits. This is also motivated by applications in health care, where a doctor's opinion is crucial in the treatment decision making process. Therefore, although contextual bandit algorithms are proven to work both theoretically and empirically in many practical settings, it is crucial to incorporate doctor's judgment to build an adaptive bandit strategy. We propose a randomized allocation strategy incorporating doctor's interventions and show that it is strongly consistent.