Harnessing probabilistic neural network with triple tree seed algorithm-based smart enterprise quantitative risk...

Area of Science:

• Artificial Intelligence
• Data Science
• Statistics
• Enterprise Risk Management

Background:

• Enterprise Management Systems (EMS) are crucial for long-term business success, integrating AI, data science, and statistics.
• Effective risk assessment within EMS is vital for informed enterprise decision-making.
• Advancements in AI, machine learning (ML), and deep learning (DL) are driving the development of sophisticated risk assessment models.

Purpose of the Study:

• To present an Improved Metaheuristics with a Deep Learning Enabled Risk Assessment Model (IMDLRA-SES) for Smart Enterprise Systems.
• To apply feature selection and deep learning for accurate business risk estimation.
• To showcase the application of applied probability and statistics in interdisciplinary studies for risk management.

Main Methods:

• Data preprocessing transforms raw financial data into a usable format.
• Oppositional Lion Swarm Optimization (OLSO) is employed for feature selection (FS) to identify optimal feature subsets.
• The Triple Tree Seed Algorithm (TTSA) optimizes a Probabilistic Neural Network (PNN) for classifying financial risks.

Main Results:

• The IMDLRA-SES technique effectively estimates business risks using feature selection and deep learning models.
• The TTSA hyperparameter optimization significantly enhances the PNN model's classification efficiency.
• Experimental evaluations on German and Australian credit datasets demonstrate superior performance over existing methods.

Conclusions:

• The IMDLRA-SES model offers a robust and accurate solution for financial risk assessment in smart enterprise systems.
• The integration of advanced metaheuristics and deep learning provides a powerful tool for improving enterprise decision-making.
• The study validates the effectiveness of applied probability and statistics in developing advanced risk management frameworks.