An Efficient Outlier Detection with Deep Learning-Based Financial Crisis Prediction Model in Big Data Environment.

Area of Science:

• Financial technology and Big Data analytics.
• Machine learning applications in finance.

Background:

• Big Data, IoT, and cloud computing are transforming business data analysis.
• Effective financial crisis prediction (FCP) is crucial for economic stability.
• Existing FCP methods lack optimal classifier efficiency and predictive accuracy.

Purpose of the Study:

• To develop an advanced model for Financial Crisis Prediction (FCP) in a Big Data environment.
• To improve the accuracy and efficiency of FCP for small and medium-sized enterprises.
• To introduce a novel feature selection algorithm for enhanced classification.

Main Methods:

• Utilized Hadoop MapReduce for Big Data management in the financial sector.
• Developed an Oppositional Ant Lion Optimizer-based Feature Selection (OALOFS) algorithm for optimal feature subset selection.
• Employed a Deep Random Vector Functional Links Network (DRVFLN) for the classification process.

Main Results:

• The OALOFS-MLC model demonstrated superior performance compared to existing approaches.
• Achieved improved classification results through optimized feature selection.
• Validated the model's effectiveness using a baseline dataset.

Conclusions:

• The OALOFS-MLC model offers a significant advancement in Financial Crisis Prediction.
• The proposed approach enhances predictive accuracy and efficiency in Big Data settings.
• This model provides a robust tool for forecasting financial failures in SMEs.