Related Work

In this chapter, we introduce the general related work for context-aware ranking with factorization models. Related work on specific issues like tag recommenders, Markov chains, etc. is discussed in detail in the corresponding chapters. Here, we discuss three general topics. The first one is recommender systems because the standard task of personalized item recommendation (a two mode problem) can be seen as context-aware ranking where the context is the user. Nevertheless in recommender systems, the term ‘context’ is usually used only for cases with at least three modes and furthermore the first mode is typically assumed to be the user. Thus, in the discussion about recommender systems we stick to the definition within the recommender community and use the term

context

 − 

aware

recommender

system only for ranking problems with at least three modes. In contrast to this, in this book we use the term

context

 − 

aware

ranking

for any number of modes. Secondly, we investigate factorization models on which our proposed approach is based. Finally, we discuss the literature about ranking in general and context-aware ranking in particular.