Application of Data Mining Method for Financial Profiling

The objective of this study is to describe financial profiles of firms via data mining method and obtain vital characteristics of performance level for the firms, which have the best financial profile of the health sector. In this study, public hospitals needed an urgent financial solution under the health sector reform steps were selected for implementation. The study covered 645 Ministry of Health hospitals, which have revolving fund in Turkey. Year of 2004 data was used in the study. As the methodology of this study, the Chi‐Square Automatic Interaction Detector or CHAID decision tree algorithm, one of the most efficient and up‐to‐date data mining method used for segmentation According to the results of the study, it was determined that financial performance of 9.15% (59 hospitals) of the covered hospitals was classified as “good” whereas 90.85% (586 hospitals) of them displayed as “bad” financial performance. Also hospitals were categorized in 12 different profiles in terms of level of financial performance by CHAID. These profiles show us what financial indicators of the hospitals should focus on for good financial performance as well as those profiles should take example to improve their financial performances. As a result of the findings, financial management policies, financial strategies and legal regulations were suggested to improve financial performance of the public hospitals and for the success of the Turkish health sector reform steps.