(DIS)UTILITY FUNCTION AND VARIABILITY EFFECT OF PARAMETER ON PREDICTION INTERVAL

• Main Author: Rizky Windrawan Farhan S, Raden
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/34230
• Institution: Institut Teknologi Bandung

Determining the value of future observation is related with prediction interval. The focus of this thesis is determining the accuracy of prediction interval. The ac- curacy of prediction interval maybe assessed by calculating its coverage probability and expected length. Besides the accuracy can be investigated by using the assess- ment of prediction interval for future observation incorporating (dis)utility function as a penalty. The prediction intervals are estimative and improved prediction intervals in which these intervals have coverage probability and expected length dier from nom- inal by O(n????1) and O(n????3=2). It is found that the penalty comes from the variability eects of parameters such as bias and mean square error of parameter. Dealing to assess the accuracy of prediction interval by using (dis)utility function as a penalty, the author used power, quadratic, and exponential (dis)utility functions.