Business Intelligence and Big Data in Credit Risk Assessment

The activities of many organizations are inextricably linked with the existence of risk. Special attention is devoted to the security of lending activities, especially in the financial services sector. Credit scoring can be carried out using a number of methods, such as descriptive methods, expert assessments, and so on. However, solutions using data mining methods are becoming increasingly recognized. They make it possible to determine the financial risk associated with individual customers. Such a process may take place at the time of concluding the contract with the customer and be based on data from their completed application forms.

Credit risk assessment can be carried out according to various models, the right choice of which depends on the purpose of the analysis and the specificity of the data being analyzed. They include the following:

■ Application scoring: completed application forms by new customers are the basis for assessing creditworthiness.

■ Behavioral scoring: additional information on customer behavior history can be used to forecast customer future behaviors.

■ Profit scoring: it takes into account not only the likelihood of a customer paying back the loan but also estimates the profit for the organization associated with cooperation with a given customer. This is a more comprehensive model because it takes into account many additional economic factors.

Thanks to the presented models, it is possible to significantly reduce the number of “bad” credits while increasing the speed of making credit decisions, which can be taken by less experienced staff. Appropriate treatment of customers who have a high risk of cessation of payments enables effective reduction of losses. The possibility of reducing the number of documents required when examining the application is also crucial.

Credit risk assessment models are used both in banking (cash loans, assessment, and delay tolerance in settling receivables), as well as in many other areas related to, e.g., renting or leasing real estate and equipment.

Business Intelligence and Big Data in Engineering and Manufacturing

In engineering and manufacturing, companies are seeking new opportunities to predict maintenance problems, enhance manufacturing quality, and manage costs through the use of BI&BD. Thanks to OLAP analyses and data mining techniques, detailed information about the costs of a technological process can be provided, which enables, among others, quick estimation of margins obtained from the sale of products at many levels of the income statement and a valuable estimation of the profitability of production of individual semifinished products in relation to alternative purchase options on the market. BI&tBD analytics also allows the analysis of the actual technical cost of production broken down into production centers, cost centers, and cost carriers (i.e., final products). It facilitates obtaining information about the material composition of products at every stage of the production process. In the same systems, nonmaterial costs allocated to products/semifinished products can be analyzed with accuracy up to the analytical items in the accounting records. The consumption rate can be calculated for individual materials and semifinished products and the unit cost in terms of the physical measure of the product for all the cost items. The BI&BD solution also enables the cost analysis in an extended cost/benefit system. This means that by analyzing the cost of manufacturing a specific product, it is possible, for example, to check what the salary costs were or what the cost of individual raw materials was in the entire multistage process. Such systems make data available for the entire company or for individual production centers. In the latter case, nonmaterial costs (remuneration, purchase, transport) incurred only in a given production center can be analyzed.

BI&BD systems allow performing simulations of the technical cost of manufacturing calculations using input values, such as volume of production, price of raw materials, consumption rate of raw and consumable materials, and value of individual nonmaterial costs that can be incurred in the period. Importantly, such simulations are carried out on the basis of real data.

With regard to product innovation, as McLaren’s Formula One cars speed around the track, for example, they send a stream of data back to the team that are processed and analyzed in real time using SAP’s Hana in-memory technology. The racing cars are laden with sensors detecting flexing, vibration, load, wear, temperature, and many other measures that impact machine performance. Hana uses sophisticated data compression to store information in random access memory, which is 10,000 times faster than hard disks, enabling the analysis of the data in seconds rather than hours. Tire real-time analysis of car sensor data is compared with historical data and predictive models, helping the team to develop improved performance, make proactive corrections, and avoid costly, dangerous incidents and, ultimately, win races. These lessons have been extrapolated to many car manufacturers who now embed their vehicles with sensors and microprocessors capturing data for maintenance and repair purposes as well as research and development innovation (Schaeffer, 2014).

< Prev   CONTENTS   Source   Next >