• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 1
  • 1
  • Tagged with
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • About
  • The Global ETD Search service is a free service for researchers to find electronic theses and dissertations. This service is provided by the Networked Digital Library of Theses and Dissertations.
    Our metadata is collected from universities around the world. If you manage a university/consortium/country archive and want to be added, details can be found on the NDLTD website.
1

交叉驗證用於迴歸樣條的模型選擇之探討

謝式斌 Unknown Date (has links)
在無母數的迴歸當中,因為原始的函數類型未知,所以常用已知特定類型的函數來近似未知的函數,而spline函數也可以用來近似未知的函數,但是要估計spline函數就需要設定節點(knots),越多的節點越能準確近似原始函數的內容,可是如果節點太多有較多的參數要估計, 就會變得比較不準確,所以選擇適合節點個數就變得很重要。 在本研究中,用交叉驗證的方式來尋找適合的節點個數, 考慮了幾種不同切割資料方式來決定訓練資料和測試資料, 並比較不同切割資料的方式下選擇節點的結果與函數估計的效果。 / In this thesis, I consider the problem of estimating an unknown regression function using spline approximation. Splines are piecewise polynomials jointed at knots. When using splines to approximate unknown functions, it is crucial to determine the number of knots and the knot locations. In this thesis, I determine the knot locations using least squares for given a given number of knots, and use cross-validation to find appropriate number of knots. I consider three methods to split the data into training data and testing data, and compare the estimation results.

Page generated in 0.0237 seconds