Subsystem Identifying Bank's Problematic Clients Among Legal Entities Based on the ML Analysis

This work reveals the concept of developing a new subsystem for identifying the bank's problematic customers based on financial indicators, news and additional information about borrowers. The structure of the subsystem, the novelty and advantages of the proposed approach are also described in detail. The description of the subject area and the study of relevant methods for the implementation of the work are given. The principles and methods of forming the factors influencing the determination of the borrower's financial condition are described. Methods for script parallel execution and data uploading and downloading organization are being developed and implemented. In addition, the implemented integration of the subsystem into the container orchestrator is described. The main purpose of the developed subsystem is to increase the efficiency and accuracy of identifying potentially insolvent borrowers.