Forecasting a Journal Impact Factor Under Missing Values Based on Machine Learning

Abstract

Scientists not only engage in research, but also write articles based on their work, and naturally aim to submit their articles to prestigious, well-received, and highly regarded journals or conferences. One way to measure the prestige of a journal is to use the Journal Impact Factor (JIF), but in order to make best use of this information one needs to better understand the relationships among various JIF-related attributes. However, in recent years JIF values have been missing or unavailable for some journals, due to several objective and subjective factors, which significantly impacts the ability to forecast such values in the following years. In this article we study the factors that directly affect the ranking of journals and can be used to forecast the appropriate values in the next few years, in order to help researchers find the right place to submit their articles in certain journals.

Keywords

• Multivariate
• time series
• missing values
• imputation
• journal impact factor
• machine learning