The procedures of transformations on non-linear data

The main purpose of this dissertation is to study the transformations on difference cases of data. There are 5 cases of data. The first case is the chemical department data which is the measurement of mangroves sediments during high tides. There are 10 of 16 independent variables of this data are n...

Full description

Saved in:
Bibliographic Details
Main Author: Lee, Cy Hau
Format: Academic Exercise
Language:English
Published: 2010
Subjects:
Online Access:https://eprints.ums.edu.my/id/eprint/13217/1/ae0000002852.pdf
https://eprints.ums.edu.my/id/eprint/13217/
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Universiti Malaysia Sabah
Language: English
Description
Summary:The main purpose of this dissertation is to study the transformations on difference cases of data. There are 5 cases of data. The first case is the chemical department data which is the measurement of mangroves sediments during high tides. There are 10 of 16 independent variables of this data are not normal and not linear. The second case of data is taken from the environment science. This data is measure about the rainfall (mm) with the factor of river water level, Gauge height, temperature, discharge and suspended solids. One dependent variable and five independent variables on this data therefore the type transformation applied is one dependent variable with many independent variables. Third case and forth case of data is taken from the forestry department. Both cases also measure about the crown of the tree diameter. The third case is the large observation of the data, n = 130 purposely used to test the effective of the transformations. The forth case data is multiple regressions which involved the combination between the variables. The last case is the measurement of the weight content within the Cigarettes. This case of data involved three independent variables and low observation size. Each of the content of the Cigarettes could be rank to see which brand is more poisonous. Normality of these cases of data allow to do further analysis such like regression analysis. Ladder and Box-Cox methods applied into these data in order to find the best power to transform the variables into more normality or linearity under difference conditions.