Phd Canditate :

E-mail :

Supervisor :

Marianna Skiada

e-mail

Assoc. Professor G. Lekakos

PhD Thesis Abstract

Recommender Systems have an enormous impact and penetration in a wide range of retail industries including businesses that promote to their customers products or services such as songs, garments, movies, gifts, accommodations, social tags etc. Nevertheless, there are demanding challenges in the field such as lack of data, data sparsity , cold start etc. that complicate the recommendation process leading to a heavy research on how το overcome each obstacle. This thesis intends to provide a road map of appropriate algorithmic solutions for businesses in ambiguous environments that face combinations of the aforementioned throwbacks. Based on the existing data conditions different recommendation strategy should be followed in order to produce accurate, interesting and useful recommendations.