Please use this identifier to cite or link to this item:
Title: Using multivariate techniques in analysing nitrous oxide emissions by continent and sector
Authors: Gauci, Stephanie (2018)
Keywords: Multivariate analysis
Cluster analysis
Nitrous oxide -- Environmental aspects
Issue Date: 2018
Citation: Gauci, S. (2018). Using multivariate techniques in analysing nitrous oxide emissions by continent and sector (Bachelor's dissertation).
Abstract: The monitoring of greenhouse gases, one of which is nitrous oxide, is crucial for the mitigation of climate change. In order to take action to reduce the emissions of such gases, First there needs to be an understanding as to what and who is causing these emissions. Three multivariate techniques, including Factor analysis, Cluster analysis and Discriminant analysis, will be used to elicit information on how different sectors within each continent are contributing to nitrous oxide emissions. Factor analysis will be used to identify patterns among the observable variables that are not clearly apparent and group correlated variables into factors which cannot be observed directly. As a result it groups observable variables into a smaller number of unobserved factors. The KMO and Bartlett's test will be used to identify whether it makes sense to use Factor analysis to elicit these latent factors. Moreover several extraction and rotation methods will be used to identify the number of latent factors and be able to interpret them. Cluster analysis is a procedure used to categorise entities into groups that are homogeneous across a range of observed characteristics. Three approaches will be used including hierarchical procedures (Ward's method), non-hierarchical procedures (K-means) and two-step clustering. The advantage of the latter procedure is that it accommodates both continuous and categorical variables and it provides the optimal number of groups. Discriminant analysis will be used to identify a discriminant function, which basically is a linear combination of quantitative predictor variables that best characterises the differences among several groups. This linear combination of predictors is similar to a multiple regression equation which combines information from several variables to separate the groups optimally.
Appears in Collections:Dissertations - FacSci - 2018
Dissertations - FacSciSOR - 2018

Files in This Item:
File Description SizeFormat 
  Restricted Access
2.35 MBAdobe PDFView/Open Request a copy

Items in OAR@UM are protected by copyright, with all rights reserved, unless otherwise indicated.