Title: Integrative and functional analysis of scientific production aspects by simplex simulation approach

Abstract

Evaluations of scientific productions are traditionally made by h-index which is defined by only the most cited publications neglecting significant part of scientists’ productions. Also, h-index can be rapidly increased by networking effects (friendship citations) leading to critical evaluation aspect of scientific productivity under reduced conception and frequent manipulation ways. This calls for bibliometric revisions through more integrative and more robust ways. In this framework, a new simplex simulation approach was developed for functional evaluations of scientific productions by highlighting regulation trends between structural variables of publications, and by considering whole sets of papers. Initially, publications are classified according to predefined criteria; then, their contents are structurally characterized by relative levels of production and cooperation variables including the numbers of pages, figures, tables, authors, affiliations, countries, etc. Using Scheffé’s mixture designs, simplex simulation iteratively combined structural variability of different publications’ classes by varying their relative weights. In response to combinations, a complete set of smoothed barycentric publication patterns was calculated. This response matrix represented system backbone from which regulation trends between production and cooperation variables were highlighted for different publications’ class. Application was illustrated by analyzing populations of Tunisian biological and medical researchers initially classified into 6 classes combining three h-index ranges with two citation levels of papers ( h or < h). Production variables showed positive trends leading to higher vs lower productivity ratios in scientists’ classes associated with lower and higher h-index ranges, respectively. Moreover, papers’ citations were improved by slight increase vs significant decrease of coauthors’ numbers, respectively. These functional results highlighted production ratios and trends that are inaccessible by h-index. By its integrative and flexible aspects, simplex simulation approach calls for developing international projects aiming for scientific productivity analyses at different scales (scientists, institutions, countries, etc.) by considering open classification criteria.

Biography

Nabil Semmar, PhD in phytochemistry (Lyon 2000), is full Professor in biological engineering at the University of Tunis El Manar (Tunisia). Since 2004, he teaches biostatistics and data mining to license, master, engineer and doctorate levels. He followed long multidisciplinary training (1988-2004) combining biological and chemical fields with computational tools (Algiers-Marseilles-Lyon Paris). In PhD, he developed a new simplex simulation method helping to highlight regulation processes governing polymorphic patterns in metabolic systems from chromatographic data. Since 2007, he published his simulation approach in many fields including plant metabolism, animal behaviors, environment assessment, pharmacology, food control, chemical synthesis and scientific production analysis. He was invited by several organisms including IAEA (Vienna, 2008) and Federal University of Parana (Brazil, 2017) to present his simulation approach. Since 2009, he edited 3 international books and 14 book chapters. In 2016, he cofounded the laboratory of bioInformatics, bioMathematics and bioStatistics in Pasteur Institute of Tunis.