Description:
the assessment of the solvency of bank client companies is influenced by many events occurring in the industry, and most of them are reflected in open news sources. Helping bank analysts analyze the diverse risks of a borrower company based on news flow involves automating the process of collecting and analyzing news by creating a solution for tagging and categorizing news according to a given set of risks.
Context:
the company has many corporate borrower clients, for whom it is necessary to constantly monitor credit risks in order to properly manage the client portfolio. One way to assess possible risks is to mention specific borrower events in the news flow (for example, a change of CEO). News flow analysis can be automated to increase efficiency.
Solution:
a news tagging model was proposed based on 17 main possible risks. Using the model, an index was built for each corporate client, demonstrating the real level of risk for all news.
Results:
Integration into the business process of the bank’s loan portfolio analysis department:
To build the model we used:
Simulation results:
A model for assessing the likelihood of risks for a given borrowing company based on previous news.
Customer: Finance, Banking
Technology stack: Python, nltk, gensim, PyTorch.
the assessment of the solvency of bank client companies is influenced by many events occurring in the industry, and most of them are reflected in open news sources. Helping bank analysts analyze the diverse risks of a borrower company based on news flow involves automating the process of collecting and analyzing news by creating a solution for tagging and categorizing news according to a given set of risks.
Context:
the company has many corporate borrower clients, for whom it is necessary to constantly monitor credit risks in order to properly manage the client portfolio. One way to assess possible risks is to mention specific borrower events in the news flow (for example, a change of CEO). News flow analysis can be automated to increase efficiency.
Solution:
a news tagging model was proposed based on 17 main possible risks. Using the model, an index was built for each corporate client, demonstrating the real level of risk for all news.
Results:
Integration into the business process of the bank’s loan portfolio analysis department:
- Automation of 60% of personnel searching for relevant information;
- Reducing the analyst’s workload by up to 70% for various risk categories;
- Increased accuracy of department forecasts by 15%;
- Increase in the number of detailed portfolio reports by 2 times;
- Compiling a semantic core for 80% of client companies.
To build the model we used:
- News flow on the selected borrowing company;
- A set of risks specified by the bank;
- Marking news according to risks from analysts.
Simulation results:
A model for assessing the likelihood of risks for a given borrowing company based on previous news.
Customer: Finance, Banking
Technology stack: Python, nltk, gensim, PyTorch.