Associate Professor
Yue Xu

Biography

Associate Prof. Yue Xu obtained her PhD in 2000. She has worked as a lecturer at Tasmania University for two years before joining in QUT as a lecturer in 2001 and later was promoted to Associate Professor at QUT in 2011.

Associate Prof. Yue Xu has been an active researcher in the areas of data analytics, machine learning, and knowledge discovery for more than 20 years. Her expertise includes pattern and association mining from large datasets, anomaly detection and vulnerability analysis, topic modelling for text data understanding and summarization, social media data analysis for improving online business and information security.

---Text mining for understanding large text corpus

With the rising and developing of Web2.0 websites, increasing amount of textual data becomes available and people need techniques to understand large text corpus quickly and accurately. Topic modeling is a useful technique to discover the theme from a collection of texts. This research is to develop new methods to evaluate the quality and interpret the semantic meaning of the discovered topics in order to correctly understand the corpus. Another focus in this area is to analyze online product reviews in order to find customers' opinions towards products. Nowadays, large number of websites encourage visitors to post reviews regarding products or services. These reviews are useful for both information publishers and readers. This sort of feedback could be used to evaluate customer satisfaction about products or services and take appropriate and sensible actions based on this information. This research is to develop text mining techniques to extract customer opinions from customer reviews.

---Modeling User Behavior on Social Networks for Detection of Social Engineering Attacks

Online social networks such as Facebook and Twitter have experienced a huge surge in popularity. The distinctive features make OSNs not only attractive to faithful users but also malicious users or attackers. Deceiving, persuading, or influencing people to provide critical information or perform an action that will benefit the attacker is known as Social Engineering. The aim of this research project is to develop data mining methods to generate user online behavior models by mining user generated online data and discover the associations between user behavior features and the source characteristics that influence users’ susceptibility to social engineering victimization.

--- Pattern and association mining

In order to help people understand and utilize massive amount of data, pattern and association mining provide techniques to automatically discover interesting patterns or associations between data items from large datasets. One big concern with the quality of association rule mining is the huge amount of discovered rules among which many are redundant thus useless in practice. Current topics in this area include:


• The development of techniques to generate compact representations of patterns and association rules in order to enhance the smart use of the discovered knowledge in practice
• The development of pattern mining techniques on large-scale heterogeneous datasets to discover cross-discipline patterns and associations
• The development of pattern based topic modelling by incorporating pattern mining techniques with statistical topic modelling to build more accurate topic models to represent a given dataset.

--- Personalized Web based Recommender Systems

Web users are often overwhelmed by the huge amount of information on the Web and are faced with the great challenge of finding the most relevant information. Recommender systems are one of the tools designed to help users deal with the information explosion by giving information recommendations. Current topics in this area include:

• The development of user profiling models (especially multi-dimensional user profiling) to discover and represent users’ information interest based on user generated content such as product ratings, tagging data, blogs, reviews, etc.
• The development of context-aware recommendation models by taking contextual information into consideration such as time, location, weather, companion, etc.
• The development of product reputation models and integrating product reputation into recommendation making
• The development of distributed recommendation algorithms to tackle the scalability problem faced by recommender systems in massive multi-dimensional data environment

Current Research Students (Principal supervisor):

• Joseph Gear, PhD candidate, Research topic: Automated Vulnerability Discovery for Industrial Control Systems using Machine Learning
• Sejung Im, PhD candidate, Research topic: Hospital readmission prediction model based on healthcare and social media data
• Yashodhya Vachila Wijesinghe, PhD candidate, Research topic: Knowledge-based Intelligenct System for automatic prediction of fall and frailty of seniors
• Dakshi Geeganage, PhD candidate, Research topic: Concept Embedded Topic Modeling Technique
• Darshika Koggalahewa, PhD candidate, Research topic: Domain Ontology learning for product recommendations based on domain information sources

Completion (Principal supervisor):

• Dr. Puspa Setia Pratiwi, PhD (2021), Thesis: A Knowledge-based Approach for Personalization in Health E-Coaching System: Case Study on Physical Activity Promotion for Individuals with Brain Impairments System Architecture for Enabling Continuous Automated Data-driven Personalization of Wellness Coaching
• Dr. Tran Diem Hanh Nguyen, PhD(2021), Thesis: Semantic-based Topic Evaluation and Application in Information Filtering
• Dr. Vin San Chia, Professional Doctorate of IT (2020), Thesis:  New Metrics for Assessing High-quality Researchers
• Dr. Bandar Alghamdi, PhD(2020), Thesis: Topic-based Feature Selection and A Hybrid Approach for Detecting Spammers on Twitter
• Anh Duc Nguyen (Dion), Master by Research (2018), Thesis: Review selection based on Topic models
• Dr. Nan Tian, PhD(2016), Thesis: Feature Taxonomy Learning from User Generated Content and Application in Review Selection
• Dr. Xiaoyu (Tyler) Tang, PhD (2016), Thesis: Recommendation Framework based on the Incorporation of Nearest Neighbourhood and Tensor Factorization
• Dr. Abdullah Ayed Algarni, PhD (2016), Thesis:  The impact of Source characteristics on users' susceptibility to social engineering victimization in social networks
• Dr. Ahmad Abdel-Hafez, PhD (2016). Thesis: Reputation Model Based On Rating Data and Application in Recommender Systems
• Dr. Yang (Grace) Gao, PhD (2015) . Thesis: Pattern-based Topic Modelling and its Application for Information Filtering
• Dr. Nora Abdullah, PhD (2012). Thesis: Integrating Collaborative Filtering and Matching-based Search for Recommending Infrequently Purchased Products
• Dr. Touhid Bhuiyan, PhD (2011). Thesis: Trust-based Automated Recommendation Making
• Dr. Huizhi(Elly) Liang, PhD (2011). Thesis: User Profiling Based on Folksonomy Information in Web 2.0 for Personalized Recommender Systems
• Dr. Gavin Shaw, PhD (2010). Thesis: Discovery and Effective Use of Quality Association Rules in Multi-Level Datasets
• Dr. Li-Tung (Soloman) Weng, PhD (2008). Thesis: Information Enrichment for Quality Recommender Systems
• Mr. Hao (Harry) Zang, Master by research (2010). Thesis: Non-redundant Sequential Association Rule Mining based on Closed Sequential Patterns
• Mr. Zhihan (John) Li, Master by research (2009). Thesis: Improvement to Chinese Information Retrieval by Incorporating Word Segmentation and Text Mining Techniques