Experiments in sparsity reduction: Using clustering in collaborative recommenders
Bridge D., Kelleher J.
© Springer-Verlag Berlin Heidelberg 2002. The high cardinality and sparsity of a collaborative recommender's dataset is a challenge to its efficiency. We generalise an existing clustering technique and apply it to a collaborative recommender's dataset to reduce cardinality and sparsity. We systematically test several variations, exploring the value of partitioning and grouping the data.