Siddharth Dinesh

Siddharth Dinesh

In this work, we have identified the need for choosing baseline approaches for research-paper recommendation systems. Following a literature survey of all research paper recommendation approaches described over the last four years, we framed criteria that makes for a well-rounded set of baselines. These are implemented on Mr. DLib a literature recommendation platform. User click data was collected as part of an ongoing experiment in collaboration with our partner Gesis. We reported the results from our evaluation for the experiments. We will be able to draw clearer conclusions as time passes. We find that a term based similarity search performs better than keyword based approaches. These results are a good starting point in finding performance improvements for related document searches.

Joeran Beel, Siddharth Dinesh

Research on recommender systems is a challenging task, as is building and operating such systems. Major challenges include non-reproducible research results, dealing with noisy data, and answering many questions such as how many recommendations to display, how often, and, of course, how to generate recommendations most effectively. In the past six years, we built three research-article recommender systems for digital libraries and reference managers, and conducted research on these systems. In this paper, we share some experiences we made during that time. Among others, we discuss the required skills to build recommender systems, and why the literature provides little help in identifying promising recommendation approaches. We explain the challenge in creating a randomization engine to run A/B tests, and how low data quality impacts the calculation of bibliometrics. We further discuss why several of our experiments delivered disappointing results, and provide statistics on how many researchers showed interest in our recommendation dataset.