1- University timetabling for maximizing the probability of
Problem: Universities usually design
timetables without any account of the behaviour of students in terms of
Objective: To analyze the attendance behavior of students from historical
data using machine learning techniques (e.g. clustering and classification) and to develop an advanced timetabling
system for universities that maximizes the probability of student attendance using mixed integer convex programming
Dr. Souad Larabi, Dr. Jaber Jamei, Dr. Ali AlMatouq, Dr. Mohammed Al-Tounisi, Dr. Dhafer AlMukhlas and Dr. Bandar AlKhayyal
Blind estimation of glucose insulin model using continuous glucose monitors
Problem: Type 1 diabetes require patient specific dynamic models that can predict the trajectory of blood glucose during meals in order to effectively maintain a healthy glucose level. Currently, developing such models is very
expensive and the life span of the model is very short (parameter uncertainties).
Objective: To obtain a glucose/insulin dynamic model of the patient using continuous glucose monitors, meal information and insulin infusion recordings for type 1 diabetes while exploiting emerging regularization techniques in system identification.
Dr. Ali AlMatouq, Dr. Mohammed AlShahrani and Dr. Taous-Meriam Laleg
Other projects to come later...