Investigating the contribution of author- and publication-specific features to scholars’ h-index prediction

EPJ Data Science

Table 2 Features used to train the machine learning models to predict the h-index

Feature group	Feature name	Description	Studies
Demographic	CareerAge	Years since first publication	[5]
	Gender	Zero for females and one for males
	MobilityScore	Number of changing the affiliation at the country level
	IncomeCurrentCountry	GDP Per Capita of current affiliation country
Prior Impact	CurrentHindex	Current h-index	[5]; [6]; [7]
	PaperPerYear	Number of total papers divided by career age	[5]; [6]; [7]
	CitationPerPaper	Number of total citations among all papers until 2008 divided by the number of all papers	[5]; [6]; [7]
Paper/Venue	PrimaryAuthorRatio	Number of papers being as primary author divided by the number of all papers
	OpenAccessRatio	Number of open access papers divided by the number of all papers
	MainField	The scientific field with the highest amount of publications
	HighRankPapersRatio	Number of publications in high-quality journals divided by the number of all papers	[5]
	DisciplineMobility	Number of unique disciplines authors has published paper divided by the number of all papers
	KeywordPopularity	Number of publications with at least one popular keyword divided by the number of all papers
	EnglishPapersRatio	Number of English papers divided by the number of all papers
Coauthor	MaxCoauthorHindex	Maximum h-index of coauthors among all papers	[15]
	CoauthorPerPaper	Number of unique coauthors among all publications divided by the number of all papers	[7]
	InternationalCoauthorRatio	Number of papers with international collaboration divided by the number of all papers