Information Gain And Entropy Methods In Banking


The paper reviews the methods of information gain and entropy used by banking companies and critically analyzes some alternatives that are possible to use.


The most common method in economics is called information acquisition, and it is based on a measure of purity called entropy (Provost and Fawcett, 2013). One of the ways to use entropy and information gain in banking is to develop analytics and work with data. This method improves the service and provides additional opportunities for infrastructure development. For example, it helps speed up the processing of big data and enhance the security of transactions in companies. Data classification is currently an established problem in information mining, which is what the entropy method helps to understand (Khare and Viswanathan, 2020). Also, thanks to this method, individual offers for customers are created. However, if the technique is used for classification, it can lead to incorrect results and require more resources, especially time. As a result, banks may experience difficulties with signs of information redundancy.

The information gain and entropy also contribute to the digitalization of payments. People actively use electronic wallets and means of charge, and cryptocurrency is developing, moving towards decentralized payment management. The interconnection of network devices and everyday human life is so difficult to monitor that it needs a reliable data mining method for information gain (Hadi and Al-Furat, 2018). Digitalization of payments is also facilitated by the spread of banking applications, which involves collecting information about the banking clientele. Moreover, this method includes using banking and payment instruments on different devices (Shetu et al., 2021). However, digital wallets and applications do not always have sufficient protection. IT specialists should consider that despite the trend toward increased security in financial technologies, digital payments continue to remain vulnerable.

One more way to use information gain is to reduce barriers to interaction with users of financial services. First of all, this includes a simple transition of the bank from one channel of interaction with the client to another. For example, entropy can be used to determine trading activity caused by news, which will reveal the mood of banking users (Liu et al., 2020). This increases the efficiency of information management and significantly simplifies the work of banking (Charbuty and Abdulazeez, 2021). Thus, the system is quite common in the modern world and has been finding practical solutions for a long time.


There are many alternatives to information gain in determining the usefulness of a variable for predicting the value of a target. For example, the gain ratio method can predict the impact of specific services on customers of the banking system (Siagian, Samsudin, and Irawan, 2022). It is also possible to use the Gini Index to solve global banking problems, such as distinguishing income inequality between countries with the same income share ratio (Sitthiyot and Holasut, 2020). The Classification and Regression Tree method, which allows for a discrete and continuous target variable, contributes to risk management in banking (Leo, Sharma, and Maddulety, 2019). Thus, the use of information gain and entropy in the IT sphere and banking has a fairly similar structure. For instance, it serves the same purpose of improving communication with customers and partners. In addition, in both cases, the presented method is the most effective compared to the alternatives and helps to make the most accurate predictions of the value of a target. The similarity also manifests itself in the fact that in both areas the presented method is much more accessible for implementation.

Reference List

Charbuty, B. and Abdulazeez, A. (2021) ‘Classification based on decision tree algorithm for machine learning’, Journal of Applied Science and Technology Trends, 2(01), pp. 20-28.

Hadi, A.A.A. and Al-Furat, A.A. (2018) ‘Performance analysis of big data intrusion detection system over random Forest algorithm’. International Journal of Applied Engineering Research, 13(2), pp. 1520-1527.

Khare, N. and Viswanathan, P. (2020) ‘Decision tree-based fraud detection mechanism by analyzing uncertain data in banking system’, Emerging research in data engineering systems and computer communications, pp. 79-90.

Leo, M., Sharma, S. and Maddulety, K. (2019) ‘Machine learning in banking risk management: A literature review’. Risks, 7(1), p. 29.

Liu, A., Chen, J., Yang, S.Y. and Hawkes, A.G. (2020) ‘The flow of information in trading: An entropy approach to market regimes’, Entropy, 22(9), p. 1064.

Provost, F. and Fawcett, T. (2013) Data Science for Business: What you need to know about data mining and data-analytic thinking. Sebastopol, CA: O’Reilly Media, Inc.

Shetu, S.F., Jahan, I., Islam, M.M., Hossain, R.A., Moon, N.N. and Nur, F.N. (2021) ‘Predicting satisfaction of online banking system in Bangladesh by machine learning’, International Conference on Artificial Intelligence and Computer Science Technology, pp. 223-228.

Siagian, S.B., Samsudin, S. and Irawan, M.D. (2022) ‘Implementation of Gain Ratio on KNN Method in Predicting Sales of Electronic Sparepart at Panasonic Service Center Lhokseumawe’. Jurnal Teknik Informatika CIT Medicom, 14(1), pp. 36-47.

Sitthiyot, T. and Holasut, K. (2020) ‘A simple method for measuring inequality’, Palgrave Communications, 6(1), pp.1-9.