Firm Clustering Based on Financial Statements

The classification and clustering of firms can help investors, consumers, and stakeholders to collect information, analyze behavior and predict potential market opportunities. Most existing approaches only pay attention to the operation processes and outputs of firms. In this paper, we introduce a firm clustering method that uses XBRL based financial information. With the adoption of XBRL, financial data has become increasingly available to allow for the application of a wide range of clustering methods. In this research, we construct the network of firms using the shared data elements in firms' financial statements. By implementing a spectral clustering method and applying it to the network, we identify clusters and characteristics of firms in the same cluster but classified as different industries according popular classification methods