With the development of technology, the world has entered a very rapid development and change process. This process has led to an increase in data for every field. For this reason, the importance of data is increasing day by day for every sector and large enterprises in many sectors reach many usable information by making use of raw data. The leading of these sectors is the banking and finance sector. Especially in the banking sector, where grouping of customers and calculating the profit obtained, especially large-volume customers, have a crucial place in order to determine marketing strategies.
This study in the banking sector consists of two stages. In the first stage, corporate customers are divided into clusters/segments according to the banking products they use. In the second stage, the profit obtained from the banking products used by corporate customers was estimated. In the study, the data of the customers in the first 6 months of 2020 were withdrawn from the database in accordance with the law on the protection of personal data. The data set consisting of corporate customers consists of 1,623 observation units and 26 variables. Observation units represent customers, and 26 variables represent banking products used by customers.
Within the scope of the study, firstly, data analysis processes were performed on the data. The K-Means method, which is one of the machine learning algorithms that gives successful results, has been applied to perform customer segmentation in the data organized by data analysis processes. With this algorithm, corporate customers using similar banking products will be included in the same clusters. Then, in order to measure the profit of the bank from each corporate customer, the profitability trend calculation was made using the machine learning algorithms K-Nearest Neighbor, Artificial Neural Networks, Random Forest and Support Vector Regression Algorithms.
Anahtar Kelimeler: Banking industry, Data mining, Machine learning, Cluster a nalysis, Artificial neural networks
|
