Laura Inozemtseva

Recommendations Systems in-the-Small

Laura Inozemtseva, Reid Holmes and Robert Walker

Published in Recommendation Systems in Software Engineering


Many recommendation systems rely on data mining to produce their recommendations. While data mining is useful, it can have significant implications for the infrastructure needed to support and maintain an RSSE; moreover, it can be computationally expensive. This chapter examines recommendation systems in-the-small (RITSs), which do not rely on data mining. Instead, they take small amounts of data from the developer's local context as input and use heuristics to generate recommendations from those data. We provide an overview of the burdens imposed by data mining and how these can be avoided by a RITS through the use of heuristics. Several examples drawn from the literature illustrate the design and applications of RITSs. We provide an introduction to the development of the heuristics typically needed by a RITS. We discuss the general limitations of RITSs.

Supplementary Material

PDF of the chapter (preprint version)