Combining multiple models to improve calibration accuracy of spectrometers

Much recent academic interest had been directed towards various multi-model ensemble techniques to produce more accurate prediction than an individual model. This holds great potential in the field of spectrometric calibration considering the vast usage of spectrometers. In this project, the author...

Full description

Saved in:
Bibliographic Details
Main Author: Tan, Jonathan Jun Wei.
Other Authors: School of Chemical and Biomedical Engineering
Format: Final Year Project
Language:English
Published: 2010
Subjects:
Online Access:http://hdl.handle.net/10356/39416
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:Much recent academic interest had been directed towards various multi-model ensemble techniques to produce more accurate prediction than an individual model. This holds great potential in the field of spectrometric calibration considering the vast usage of spectrometers. In this project, the author used the bagging and boosting technique as committee machines to complement the Partial Least Squares Regression and Gaussian Process Regression methodologies in the calibration of the "Tablets" and "Meat" dataset. The results indicate their superiority over single model prediction and upon comparing the bagging and boosting algorithm on a single dataset, it appears that the boosting technique is marginally better.