Outlier detection in cylindrical data / Nurul Hidayah Sadikon

Nurul Hidayah , Sadikon (2018) Outlier detection in cylindrical data / Nurul Hidayah Sadikon. Masters thesis, University of Malaya.

[img] PDF (The Candidate's Agreement)
Restricted to Repository staff only

Download (1521Kb) | Request a copy
    [img]
    Preview
    PDF (Thesis M.A)
    Download (937Kb) | Preview

      Abstract

      A cylindrical data set consists of a circular and a linear variables. Few distributions have been proposed for such data pioneered by Johnson and Wehrly (1978). In this study, we look at two problems of detecting outliers in cylindrical data. Firstly, we define outlier in cylindrical data and propose a new test of discordancy to detect outlier in cylindrical data generated from Johnson-Wehrly distribution. Secondly, we focus on detecting outliers in Johnson-Wehrly circular-linear regression model. In both cases, the outlier detection procedures are developed using the k-nearest neighbor distance. The cut-off points are obtained and the performance of the new statistic is examined via simulation. A practical example is presented using the wind data set from the Malaysian Meteorological Department. The findings of the study should lead to better inferences, model fitting and forecasting of cylindrical data sets.

      Item Type: Thesis (Masters)
      Additional Information: Dissertation (M.A.) – Faculty of Science, University of Malaya, 2018.
      Uncontrolled Keywords: Cylindrical data; Cylindrical regression model; Detection procedure; K-nearest neighbor’s method; Outlier
      Subjects: Q Science > Q Science (General)
      Q Science > QA Mathematics
      Divisions: Faculty of Science
      Depositing User: Mr Mohd Safri Tahir
      Date Deposited: 13 Dec 2018 02:26
      Last Modified: 13 Dec 2018 02:26
      URI: http://studentsrepo.um.edu.my/id/eprint/9371

      Actions (For repository staff only : Login required)

      View Item