Modelling of Factors Influencing the Citation Counts in Statistics

Citation is considered as the most popular quality assessment metric for scientific papers, and it thus important to determine what factors promote citation count of a paper in comparison others same field. The main aim this study model counts research published SCI or SCI-Expanded journals Statistics field with growing number works Turkey. It well known that right-skewed nature makes classical regression modelling inappropriate, even if log transformation applied [1]. Due fact distribution involves great zeros, serves an additional advanced discrete models more precise prediction [2]. Data collected consist all papers produced by 261 Statisticians between 2005-2017. Discrete varying from Poisson Zero-Inflated Hurdle were constructed possible influential factors, such publication age, references, journal category etc. Predictive performances alternative compared via AIC Vuong test [3]. Results suggested Zero Inflated Negative Binomial mixture are best forms predict zero inflation [4]. In addition, final interpreted make some suggestions increase their work.