Last update: May 2024
Stats Savvy. Tech Trailblazer. Linguistic Luminary.
Wai Hong Tan, Ph.D.
Doctor of Philosophy in Mathematics, UNSW Sydney
I obtained my Ph.D. in Mathematics from UNSW Sydney in 3.5 years under full sponsorship by the Ministry of Higher Education, Malaysia. My research interest revolves around applying the theory of point processes to model a diverse array of real-life phenomena. My tagline “Stats Savvy. Tech Trailblazer. Linguistic Luminary.” essentially highlights my areas of expertise.
Stats Savvy – Completing my Ph.D. in Mathematics (with a focus on Applied Statistics) speaks volumes about my statistical expertise. I have published several papers in prestigious journals, been invited by high-impact journals to review articles, and delivered plenary speeches. I have also received best paper awards during conferences.
Tech Trailblazer – My expertise in computer science converges seamlessly with my statistical acumen, earning me positions as a facilitator of various workshops for both undergraduate and postgraduate levels. Proficient in navigating diverse technological landscapes, I stay current by ensuring effective utilization of cutting-edge tools and methodologies.
Linguistic Luminary – My linguistic prowess has earned recognition from reputable institutions such as the University of Cambridge Local Examination Syndicate, which awarded me a Distinction in English Language. Proficient in five different languages, I have undertaken numerous crucial roles, such as serving as the editor for documentation related to the U.S. Embassy Grant and offering live translation services to international students.
Educational Background
Doctoral Degree
The University of New South Wales (UNSW)
I completed my Doctoral degree under 3.5 years of time. My PhD thesis is entitled "Predicting the Popularity of Tweets Using the Theory of Point Processes". UNSW is ranked 19th in the world, based on QS World University Rankings 2024.
Certified by My eQuals Australia,
the official tertiary credentials platform.
Bachelor's Degree
The Northern University of Malaysia (UUM)
I completed my Bachelor's degree under 3 years of time, being conferred the Bachelor of Decision Science (with Honors), with a cumulative grade points average of 3.65. UUM is ranked 538th in the world, based on QS World University Rankings 2024.
Recent Works
Predicting the Popularity of Tweets Using the Theory of Point Processes
PhD Thesis
This thesis focuses on the problem of predicting the tweet popularity, or the number of retweets stemming from an original tweet. We propose several prediction methodologies using the theory of point processes, where the prediction of the future popularity of a tweet is based on observing the retweet time sequence up to a certain censoring time, and the prediction performance is evaluated on a large Twitter data set.
This thesis contains seven different chapters, with Chapter 1-3 being more of introductory accent, Chapter 4-6 being the main body of the thesis, and Chapter 7 being the concluding remarks.
Marked Self-Exciting Point Process Modelling of Information Diffusion on Twitter
Published Manuscript
We propose a reliable tweet popularity prediction approach based on a marked self-exciting point process model, motivated by the observation that retweet activities tend to occur in clusters or bursts. Using suitable prediction functionals, we demonstrate that the proposed approach is capable of predicting the future popularity levels of tweets more accurately than those based on the existing approaches, especially at shorter censoring times.
The manuscript has been published in the Annals of Applied Statistics, and its content has been included as Chapter 4 in the main thesis.
Predicting the Popularity of Tweets Using Internal and External Knowledge: An Empirical Bayes Type Approach
Published Manuscript
We propose a novel empirical Bayes type approach to combine knowledge learned from the complete retweet time sequences in the training data with that currently observed in the test data to estimate the parameters of different models. Using suitable prediction functionals, we highlight that point process models applying the approach exhibit superior prediction performances compared to their counterparts with parameters conventionally estimated.
The manuscript has been published in AStA Advances in Statistical Analysis, and its content has been included as Chapters 5 and 6 in the main thesis.
On the Choice of Functionals Obtained from the Predictive
Distribution of Future Retweet Counts
Published Manuscript
We propose order (–1) and harmonic medians as theoretically optimal functionals relative to the mean and median absolute percentage errors respectively, being two of the most prominent metrics used in assessing the accuracy of tweet popularity prediction. We outline how the two functionals can be obtained for a relatively simple Poisson process model and other more complex predictive models, highlighting the issues pertaining to optimality in the process.
This manuscript has been presented in the 3rd International Conference on Applied & Industrial Mathematics and Statistics 2022 (ICoAIMS 2022) and has been published in AIP Conference Proceedings. Its content has been included as Appendix A in the main thesis.
Appointed as a joint researcher in international grants
such as the U.S. Embassy and Matching grants
Invited as a reviewer of high-impact journals
such as the Journal of Supercomputing and Information & Management
Nominated as a plenary speaker in prestigious events
such as UNSW Sydney Statistical Conference
Recognized for English Language proficiency
such as a Distinction in the University of Cambridge Local Examinations Syndicate Assessment
Served as a language referral
such as in U.S. Embassy grant documentation and PD/viva sessions for postgraduate students
Acquired several best paper awards
such as in ISEBT2021 and ICoAIMS2022
Involved in projects with income-generation amounting to 150,000$
such as ISEBT2021, InCEBT2022, and InCEBT2023