Title of Project
Mathematical and Computational Tools for Big Data Analytics and Applications
Prof. Ricardo J. G. B. Campello
College or Research Centre
College of Science & Engineering
Summary of Project
Terms such as data mining, big data and data science have become buzzwords virtually in all fields of knowledge, as our capacity to acquire data from various sources has far surpassed our human ability to analyse and extract knowledge from the resulting databases. Nevertheless, not all types of data can be easily acquired. Indeed, unlike unlabelled observations, which can often be measured in a cheap, fully automatized way, labelled observations can be very laborious and expensive to obtain, as they typically require human intervention (e.g. medical diagnoses). This gives rise to an increasing gap between the superabundance of unlabelled data and the scarcity of labelled data in many domains. This project focuses on the development of new mathematical and algorithmic tools for data mining in fully or partially unsupervised application scenarios, where none or only a small fraction of labelled data is available. Other aspects and challenges of big data are also in the spotlight, specially the efficient processing of large volumes of data as well as the processing of high-dimensional datasets. Particular emphasis in this project is given to the tasks of data clustering, outlier detection, semi-supervised classification, and their applications.
Data Science; Data Mining; Big Data; Clustering; Outlier Detection; Semi-Supervised Classicitaion; Applications
Would suit an applicant who
Is passionate about algorithms, discrete maths/optimization, and/or applied statistics.
Updated: 10 months ago