Formulation of rules for the scientific community using deep learning

In a deluge of scientific literature, it is important to build scientific quantitative rules (SQR) that can be applied to researchers' quantitative data in order to produce a uniform format for making decisions regarding the nomination of outstanding researchers. Google Scholar and other search engines track scholars' papers, citations, etc. However, the scientific community hasn't agreed on standards a researcher must meet to be regarded as important. In this paper, we suggest rules for the scientific community based on the top five quantitative scientific parameters. The significance of the parameters is measured based on two factors: (i) parameters' impact on the model's performance while classifying awardees and non-awardees, and (ii) the number of award-winning researchers elevated in the ranking of researchers through each respective parameter. The experimental dataset includes information from researchers in the civil engineering, mathematics, and neuroscience domains. There are 250 awardees and 250 non-awardees from each field. The SQR for each discipline has attained an accuracy of 70% or more for their respective award-winning researchers. In addition to this, the top ranked parameters from each discipline have elevated more than 50% of the award-winning researchers into their respective ranked lists of the top 100 researchers. These findings can guide individual researchers to be on the list of prestigious scientists, and scientific societies can use the SQR to filter the list of researchers for subjective evaluation in order to reward prolific researchers in the domain.

Automated summary

This paper proposes a scientific algorithm to establish a uniform standard for nominating outstanding researchers based on quantitative data analysis. The effectiveness of the algorithm is demonstrated through experimental data analysis.