OpenHarvesters harvest publications metadata from different open source databases (CrossRef, PubMed and DBLP) and identify publications for a scholar. Name of the author can be mentioned differently, in the citation data of a publication of a source. For example, "Dean Blackmar Krafft", "Dean B. Krafft", "Dean Krafft", "DB Krafft", "D. Krafft" etc. These name variations make it harder to identify publications for a scholar using a named "search string". The difference between two distinct author names could be just the middle name initial e.g., "David F. Stern" and "David B. Stern". OpenHarvester algorithm learns from the claim publications as well as from the existing citation data of a claimed article. Currently, claimed publications can be stored in CSV, TXT, PDF and VIVO-Model format. No work is currently done in regards to "performance improvement" but focus was on "identification of correct publications".
Please see description on https://github.com/mjaved495/OpenHarvester