上海财经大学 > 科学研究 > 学术交流 > 学术报告
  • 统计与管理学院2018年学术报告第2期

    【主 题】 kNN estimation in functional partial linear modeling

    【报告人】 凌能祥 教授

    合肥工业大学数学科学学院

    【时 间】 2018年01月08日(星期一)09:30-10:30

    【地 点】 上海财经大学统计与管理学院大楼1312会议室

    摘 要】A statistical procedure combining the local adaoptivity and the easiness of implementation of k-nearest-neighbours (kNN) estimates together with the semiparametric flexibility of partial linear modeling is developed for regression problems involving functional variable. Various asymptotic results are stated, both for the linear parameters and for the nonparametric operator involved in the model. A simulation study compares the finite sample behaviour of the kNN method with alternative estimation procedures. Finally, comparison with alternative functional regression models is carried out by means of a real curves data application which exhibits the interest both of the k NN method and of the semi-parametric modeling.

    嘉宾简介】合肥工业大学数学学院教授,硕士生导师,美国《数学评论》特约评论员。研究方向:非参数统计; 近代回归分析; Bayes 统计。