Analysing and identifying website personality by extending existing libraries
Loading...
Other Title
Authors
Chishti, Shafquat Ali
Author ORCID Profiles (clickable)
Degree
Master of Computing
Grantor
Unitec Institute of Technology
Date
2016
Supervisors
Li, Xiaosong
Sarrafzadeh, Hossein
Sarrafzadeh, Hossein
Type
Masters Thesis
Ngā Upoko Tukutuku (Māori subject headings)
Keyword
website personalities
websites
evaluation
Website Personality Scale
library (computing)
automated evaluation
websites
evaluation
Website Personality Scale
library (computing)
automated evaluation
ANZSRC Field of Research Code (2020)
Citation
Chishti, S.A. (2016). Analysing and Identifying Website Personality by Extending Existing Libraries. An unpublished thesis submitted in fulfilment of the requirements for the degree of Masters in Computing,
Unitec Institute of Technology, New Zealand.
Abstract
Previously, website personality was only assessed and classified by human interaction. This brings with it a host of problems as humans act depending on their likes and dislikes. For example, if someone likes the particular colour of a website he will classify it as attractive but if he does not like the particular colour he will deem the web site unattractive. To remove these sorts of problems that come with the aspect of human bias, an impartial decision maker is needed. As every living thing that has a mind of its own will have some biases, a machine, more specifically a computer, is the best option. A computer can analyse and categorise website personality on the basis of quantitative elements of the website. Hence, a software tool needs to be developed to assess and classify website personality.
Experiment has been carried out for the research using a software tool. The software tool that I have developed is designed to work on the same lines as the Website Personality Scale research done by human beings, which involved classification of website personalities by research and surveys. The only difference is that of the human bias, which is removed by using the software tool. K-means algorithm is used in the tool to classify a website on the basis of the data collected from website pages. To train the software tool a website data bank was made which contained 240 websites; 112 new websites were tested on the developed software tool, with positive results showing how close results from a test website are to the training websites. The tool can successfully identify and analyse a website and classify it with similar training websites from the data bank. The whole process is fast and automatic without the need for any human involvement.
Publisher
Permanent link
Link to ePress publication
DOI
Copyright holder
Author
Copyright notice
All rights reserved