اراﺋﻪ ﻣﺪل ارزﯾﺎﺑﯽ ﻋﻤﻠﮑﺮد ﺑﺎﻧﮏﻫﺎي ﺑﻮرﺳﯽ ﮐﺸﻮر ﺑﺎ رﻫﯿﺎﻓﺖ داده ﮐﺎوي

Presenting a Model Based on Evaluation of Performance Banks Listed in Tehran Stock Exchange Using Data Mining Approach

ﺑﺎ رﺷﺪ ﺑﺎﻧﮏﻫﺎي ﺧﺼﻮﺻﯽ و ﻣﺆﺳﺴﺎت ﻣﺎﻟﯽ و اﻋﺘﺒﺎري،رﻗﺎﺑﺖ ﺑﯿﻦ آﻧﻬﺎ ﺑﻪ ﻣﻨﻈﻮر اراﺋﻪ ﺧﺪﻣﺎت ﺑﻬﺘﺮ اﻓﺰاﯾﺶ ﯾﺎﻓﺘﻪ اﺳﺖ. ﺑﺎ ﺗﻮﺟﻪ ﺑﻪ اﻫﻤﯿﺖ ﻣﻮﺿﻮع، ﺗﺪوﯾﻦ ﯾﮏ اﻟﮕﻮي ﺟﺎﻣﻊ و ﮐﺎﻣﻞ ﺑﺮاي ارزﯾﺎﺑﯽ ﺑﺎﻧﮏﻫﺎ ﺿﺮوري ﺑﻪ ﻧﻈﺮ ﻣﯽرﺳﺪ. ﻫﺮ ﺳﺎزﻣﺎن ﺑﺮاي آﮔﺎﻫﯽ از ﻧﻘﺎط ﻗﻮت و ﺿﻌﻒ ﺧﻮد ﺑﻪ وﯾﮋه در ﻣﺤﯿﻂﻫﺎي ﭘﻮﯾﺎ ﻧﯿﺎزﻣﻨﺪ ارزﯾﺎﺑﯽ ﻋﻤﻠﮑﺮد ﻣﯽﺑﺎﺷﺪ. ﻣﻮﺿﻮع ارزﯾﺎﺑﯽ ﻋﻤﻠﮑﺮد آﻧﻘﺪر ﻣﻮرد ﺗﻮﺟﻪ اﺳﺖ ﮐﻪ ﺣﺘﯽ در اﯾﻦ ﺑﺎره ﺻﺎﺣﺐﻧﻈﺮان ﻣﺪﯾﺮﯾﺖ ﻣﻌﺘﻘﺪﻧﺪ: آﻧﭽﻪ را ﮐﻪ ﻧﺘﻮان ارزﯾﺎﺑﯽ ﻧﻤﻮد، ﻧﻤﯽﺗﻮان ﻣﺪﯾﺮﯾﺖ ﮐﺮد". ﺑﻨﺎﺑﺮاﯾﻦ ﺑﺎﻧﮏﻫﺎ ﻧﯿﺰ ﻫﻤﺎﻧﻨﺪ ﺳﺎﯾﺮﺳﺎزﻣﺎنﻫﺎ در اﯾﺮان ﺑﺮاي اراﺋﻪ ﺧﺪﻣﺎت ﻣﺘﻨﻮعﺗﺮ و ﺳﺮﯾﻊﺗﺮ و ﻫﻤﭽﻨﯿﻦ ﺗﻮﺳﻌﻪ ﺧﻮد ﻧﯿﺎزﻣﻨﺪ ارزﯾﺎﺑﯽ ﻋﻤﻠﮑﺮد ﻣﯽﺑﺎﺷﻨﺪ.]6[ ﻫﺪف اﺻﻠﯽ اﯾﻦ ﻣﻘﺎﻟﻪ اراﺋﻪ ﻣﺪﻟﯽ ﺑﻪ ﻣﻨﻈﻮر ارزﯾﺎﺑﯽ ﻋﻤﻠﮑﺮد ﺑﺎﻧﮏﻫﺎي ﺑﻮرﺳﯽ ﮐﺸﻮر ﺑﺎ اﺳﺘﻔﺎده از روشﻫﺎي داده ﮐﺎوي ﻣﯽﺑﺎﺷﺪ. در اﯾﻦ ﭘﮋوﻫﺶ، 4 ﻣﺪل داده ﮐﺎوي درﺧﺖ ﺗﺼﻤﯿﻢ C5.0، درﺧﺖ ﺗﺼﻤﯿﻢ C4.5، اﻟﮕﻮرﯾﺘﻢ ﻧﺎﯾﻮ ﺑﯿﺰ و ﺟﻨﮕﻞ ﺗﺼﺎدﻓﯽ، ﺑﻪ ﻣﻨﻈﻮر ارزﯾﺎﺑﯽ ﻋﻤﻠﮑﺮد ﺑﺎﻧﮏﻫﺎي ﭘﺬﯾﺮﻓﺘﻪ ﺷﺪه در ﺑﻮرس اوراق ﺑﻬﺎدار ﺗﻬﺮان ﭘﯿﺎدهﺳﺎزي و ﺑﺎ ﯾﮑﺪﯾﮕﺮ ﻣﻘﺎﯾﺴﻪ ﻣﯽﮔﺮدﻧﺪ. ﺑﺪﯾﻦ ﻣﻨﻈﻮر ﻧﻤﻮﻧﻪاي ﻣﺸﺘﻤﻞ ﺑﺮ 28 ﻧﺴﺒﺖ ﻣﺎﻟﯽ ﺷﺎﻣﻞ ﻧﺴﺒﺖﻫﺎي ﺳﻮدآوري، ﻧﻘﺪﯾﻨﮕﯽ، ﺣﻮزه ﮐﯿﻔﯿﺖ ﻣﺪﯾﺮﯾﺖ ، ﺣﻮزه ﮐﯿﻔﯿﺖ داراﯾﯽ و ﮐﻔﺎﯾﺖ ﺳﺮﻣﺎﯾﻪ در 18 ﺑﺎﻧﮏ ﭘﺬﯾﺮﻓﺘﻪ ﺷﺪه در ﺑﻮرس و ﻓﺮاﺑﻮرس اوراق ﺑﻬﺎدار ﺗﻬﺮان درﻓﺎﺻﻠﻪ ﺑﯿﻦ ﺳﺎﻟﻬﺎي 1393 ﺗﺎ 1396 ﺑﻪ ﻋﻨﻮان ﻣﺘﻐﯿﺮﻫﺎي ﻣﺴﺘﻘﻞ و ﻋﻤﻠﮑﺮد ﺑﺎﻧﮏﻫﺎ در ﺳﻪ دﺳﺘﻪ ﻗﺎﺑﻞ ﻗﺒﻮل، ﻏﯿﺮﻗﺎﺑﻞ ﻗﺒﻮل وﻣﺘﻮﺳﻂ ﺑﻪ ﻋﻨﻮان ﻣﺘﻐﯿﺮ واﺑﺴﺘﻪ اﻧﺘﺨﺎب ﮔﺮدﯾﺪﻧﺪ. ﻧﺘﺎﯾﺞ ﺣﺎﮐﯽ از آن اﺳﺖ ﮐﻪ در ﺑﯿﻦ ﮐﻠﯿﻪ ﻣﺪلﻫﺎ درﺧﺖ ﺗﺼﻤﯿﻢ C5.0ﺑﺎ ﺻﺤﺖ 94,4% ﺑﻬﺘﺮﯾﻦ ﻣﺪل اراﺋﻪ ﺷﺪه در اﯾﻦ ﭘﮋوﻫﺶ ﻣﯽﺑﺎﺷﺪ.

With the growth of private banks , financial and credit institutions, competition for better services has increased. Given the importance of the issue, it is necessary to develop a comprehensive model for evaluating banks. Every organization needs to evaluate its performance to understand its strengths and weaknesses, especially in dynamic environments. The issue of performance appraisal is so widespread that even management experts say: "What cannot be evaluated cannot be managed". Banks, like other organizations in Iran, need performance evaluation to provide more diverse and faster services as well as their development. [6] This study aimed to present a model to evaluate the performance of banks listed in Tehran Stock Exchange using data mining approach. In this research, four data mining models of decision tree C5.0, decision tree C4.5, Naive Bayes classifier, and random forest were implemented and compared to evaluat the performance of banks. To this end, 28 financial ratios (e.g., profitability ratios, liquidity, quality management, asset quality, and capital adequacy) in 18 banks of Tehran Stock Exchange during 2014-2017 were selected as independent variables. In addition, the performance of banks in three categories of acceptable, unacceptable, and moderate was selected as the dependent variable of the study. According to the results, the decision tree C5.0 with the accuracy of 94.4% was the most efficient model proposed in this research.