What to Read? With Whom to Work? Where to Publish? - Scientific Techniques for Organizing and Conducting Engineering Research

There are an ever more rapidly increasing number of design and CAD conference and journal papers. In particular, there are also a large number of design researchers (paper authors) and a large number of microelectronics conferences. Significant government and industrial funding and even higher research effort is allocated to a great variety of topics. While some researchers keep changing collaborators, others maintain steady collaborations. The return on financial and human investment is sometimes spectacular but often minimal or none. Our conceptual aim is to help microelectronics researchers, students, designers, and managers better organize their professional activities. Our practical objective is to address some of the essential productivity questions for microelectronics researchers, including what papers to read, where to publish, and with whom to collaborate. Our scientific goal is to develop statistically sound technique that guide actions of highly successful microelectronics researchers, which we define for the purpose of this paper as ones that are the authors of the most cited papers. We show that one can surprisingly accurately predict the number of citations for a paper at the time of its publishing by considering factors such as citation factors of the authors and the quality of a conference or a journal. Interestingly, highly cited coauthors rarely improve the citation record of their collaborators. Keywords-Citation Prediction; Machine Learning; Bibliometrics