The number of IoT malware variants infecting vulnerable IoT devices is growing. Many of these are created by modifying parts of publicly available source codes and adding, changing, or deleting functions. In a previous study, we defined a Function Call Sequence Graph (FCSG) as a directed graph representing the sequence of functions that a specimen calls. Then, we found the common functionality among specimens of malware variants. In this paper, we associate this method with IoT malware clustering outcomes, extracting representative subgraphs of FCSGs extracted from the elements of each cluster. We call each of the subgraphs as a representative FCSG (R-FCSG). We calculate the similarity between R-FCSGs to identify clusters exhibiting similar or identical functionality across CPU architectures. We also studied a method to resolve the family name of clusters by focusing on the similarity between R-FCSGs with no determined family name and R-FCSGs with family names.