ITx 2020 Programme

Jump to: Wednesday 7th | Thursday 8th

Using a mobile device?
Visit programme on mobile site

Please note as with our physical conferences, all of the virtual conference sessions will be recorded.

To connect to the presentations please see the link in each of the presentation descriptions in the programme or use the following stream presentation links

 

For the CITRENZ1 StreamFor the CITRENZ2 StreamFor the CITRENZ3 Stream
http://itp.nz/citrenz1 http://itp.nz/citrenz2 http://itp.nz/citrenz3 

This years poster presentations can be viewed at itp.nz/citrenzposters

Voting on the posters is also being held online, vote at itp.nz/citrenzvoting

Improved Feature Selection and Ensemble Learning For Cervical Cancer Assessment

Wednesday 12:00pm - 12:30pm, CITRENZ (https://itp.nz/citrenz3)

Choosing the right influencing feature is a challenging field in data science due to the presence and complexity of multi-dimensional data. Cervical cancer is an excellent example for such study, as well as impacting individuals and families, presents almost no-symptoms at the early stages of development of this condition. Because multi-factors may be involved, this demands a lot of research and analysis to identify causative or linked features. The researchers have applied and optimised an ensemble learning algorithm as it is the best model for multi-modal medical data when relatively high dimensionality is present. The main objective of this study was to minimize the dependency on data pre-processing techniques, whilst analysing the data (filling/ignoring missing values with the statistical method). Main factors were studied and validated using Root Mean Square Error (RSME) and Mean Absolute Error (MAE).
The classification accuracy for features were obtained by 10-fold cross-validation and test (where 66% is training data and 34% test data). The data was obtained from the UCI machine learning repository. WEKA and MATLAB were used to identify features. SPSS and SAS were used for RMSE and MAE. This approach is generic, and may also be applied to any relevant dataset for other purposes, and for teaching data analytics.
Keywords: Feature selection, ensemble learning, data mining, machine learning, models, HPV, WEKA, MATLAB.

Speakers

Rajib Hasan

School of Computing, AUT

Rajib has earnt a 1st MPhil majoring machine learning, 2nd M. Phil majoring data science, and a bachelor's degree in artificial intelligence majoring. He has secured several international software research and innovation awards as well as several copyrighted software. Rajib is an active researcher on data science and experienced cyber security professional. Learning and sharing the knowledge through tertiary teaching.

Noor Alani

Eastern Institute of Technology

Noor is an academic staff member at the School of Computing – EIT. He obtained Ph.D. from AUT in 2019 specialised in computer vision models to improve traffic and pedestrian safety. His research interests include machine vision, emerging technologies, and data analytics. In 2016 he had a visiting position at Wuhan University, China. He received numerous awards including the Akria Nakamura awards for 2017, and 2018 from AUT, the “Best Student Paper Award” at CAIP 2017 – Springer, and best presentation award from ICCAR 2018 - IEEE.