An Application of the Bipolar Metric Space Structure for Missing Data Imputation in Ligand–Target Interactions

In this study, the bipolar metric space structure is tested on a real dataset for the first time. In this context, ligand–target interaction scores are examined and transformed into a form compatible with the bipolar metric function, an incomplete-data scenario is evaluated, and in addition, the closest pairs of the same type are identified and the resulting findings are analyzed. The model is also applied to provide theoretical estimations for some missing entries in the dataset. By emphasizing that the proposed method can be tested under various topics in both the natural sciences and the humanities, the study thus presents a pioneering example in this direction.