Recommendation Engine for Master Thesis based on Lenskit .
This work pursues to study the impact of different neighborhood strategies in the formulation of recommendations through collaborative methods. Several popularity strategies are studied and implemented using the framework Lenskit . The aim is to show the usefulness of popularity as a significant indication in the creation of recommendations.
The different algorithm configuration files can be find at etc directory, where we set every parameter required. You can read more about it here .
The first changes focused on parallelizing the way similarities are computed. Due to some issues related with the size of datasets, several stages of the job had to be serialized as well.
