Data Mining
Data mining, or DM, is the process of locating data that has been dispersed within a sizable database. It is a new field that incorporates various specialties, such as database management, statistics, and information science. It is utilized to extract trends and patterns from the recorded data that human analysis would not reveal (Moin & Ahmed, 2012; Pulakkazhy & Balan, 2013). Data mining uses a variety of methodologies.
Association Mining
Items whose relationship or connection had not yet been established can be correlated via association. A big data set is mined using this data mining technique to uncover hidden relationships. If a relationship of data points satisfies predetermined confidence levels, it is considered significant. The associations generated are subjected further to statistical analysis to evaluate the type and significance of association. It outlines data items which tend to occur together hence used in the market basket analysis to determines which purchases are made together in order develop a marketing strategy for profit generation (Moin & Ahmed, 2012; Pulakkazhy & Balan, 2013).
Classification and Prediction
Classification and prediction is another DM technique which is utilized when the classes of data are already known. The data in the known classes is used to develop models that are used to predict classes for the data whose classes are unknown. The predicting model or the classifier is developed using a decision tree or a neural network tree model. Classification techniques work with discrete data which are also unordered to identify labels for different classes. Prediction techniques use valued functions such as regression analysis to predict missing values in unknown data. Clustering analysis is another DM technique. It is similar to classification except that the classes of data are unknown. Data items are classified into groups using the principle of maximizing the similarity to an observed pattern (Moin & Ahmed, 2012; Pulakkazhy & Balan, 2013).
DM in the Banking Sector
DM in the banking sector can be used to predict customer behavior based various customer details already present in the database. The information can be used to profile customers and segregate them based on their preferences. This would be a great aid especially when deciding on introducing a new product into the market, i.e. targeted marketing. Those customers whose profiles are in line with the product to be launched are highly targeted, those whose profile indicate not likely to be interested in a product are left out. This saves on resources as such customers could have been contacted in the absence of data mining and their non-participation could not return resources used in operation. This profiling also plays a role in risk management. Based on the data, an institution can identify credit worthiness of customers requesting for loans thus knowing their chances of defaulting or repaying the loan (Moin & Ahmed, 2012; Pulakkazhy & Balan, 2013). The transaction patterns can be analyzed and any deviation from the identified patterns alerts the bank to check for fraud. This implies that any fraudulent transactions could be detected and arrested before it affects the profitability of financial institutions. Regression models are applied to establish the presence of linear or nonlinear relationships. In predicting stock market prices, regression analysis works with the current prices and the prevailing market conditions to project the price at a given time (Moin & Ahmed, 2012; Pulakkazhy & Balan, 2013).
Kmeans Algorithm
Kmeans algorithm is used in clustering analysis.
Regression Models
Regression models are used for prediction while Bayesian algorithms are used for classification.
These algorithms are important for my financial institution as they would play a big role in profiling the customers to determine their behavior and also for selective marketing. Predictive algorithms will help the bank avert losses due to default loan repayments and due to fraudulent transactions.
References
Moin, I. K., & Ahmed, B. Q. (2012). Use of Data Mining in Banking. International Journal of Engineering Research and Applications, 2(2), 1-5. Retrieved from http://ijera.com/papers/Vol2_issue2/DU22738742.pdfPulakkazhy, S., & Balan, R. V. S. (2013). DATA MINING IN BANKING AND ITS APPLICATIONS-A REVIEW. Journal of Computer Science Published Online, 9(910), 1252-1259. https://doi.org/10.3844/jcssp.2013.1252.1259
