The representative works of scientists

Nowadays identifying the personal representative works is becoming increasingly important and necessary for scientists in many cases, such as faculty hiring and promotion applications. There are already a few methods based on different criteria for selecting the representative works of a scientist, like citation count. In addition, we can observe that some researchers always produce many similar quality scientific papers and some researchers have several highly cited papers compared with his or her other papers. In this context, we propose to use the maximum gap in a histogram of a scientist’s sorted papers’ citation counts to classify his or her papers into two groups, i.e. representative papers and regular papers. Based on the maximum gap, we then design an indicator (D_{r}) to quantify the impact difference between scientist’s representative works and regular works. We apply this selection method and (D_{r}) index into the data of American Physical Society (APS) journals. The results indicate that the selection method can better identify the representative works of Nobel laureates in Physics compared with using the most cited paper. We also find that the number of representative works selected by our method is related to (D_{r}). A larger number of selected papers would appear when the value of (D_{r}) index is relatively smaller. Meanwhile, we also observe that (D_{r}) is weakly correlated with the h index and total citation.