Combining self-organizing map and K-means clustering for detecting fraudulent financial statements

Auditing practices nowadays have to cope with an increasing number of fraudulent financial statements (FFS). Based on data mining techniques, researchers have made some studies and have found that the techniques can facilitate auditors in accomplishing the task of detection of FFS. However, most of the techniques used in the detection of FFS are supervised methods. Clustering, one kind of unsupervised data mining technique, has almost never been used. Therefore, considering the characteristics of FFS and self-organizing map(SOM), a model combining SOM and K-means clustering based on a clustering validity measure is designed. To carry out the experiment, 100 financial statements from Chinese listed companies during 1999-2006 are selected as experimental sample according to some specific standards. 47 financial ratios are chosen as variables. The model is applied to the data and good experimental results are obtained.