LEP-QNN: Loan Eligibility Prediction using Quantum Neural Networks

TL;DR

The paper tackles loan eligibility prediction in finance by introducing LEP-QNN, a quantum neural network that uses angle encoding, a fixed-depth ansatz, and dropout for regularization. It conducts a comprehensive optimizer and noise-model analysis, finding Adam to deliver the best performance with 98% accuracy on a real-world dataset. The work also demonstrates the model’s resilience to several quantum noise models, and it benchmarks LEP-QNN against classical ensemble methods, showing a significant accuracy advantage. Collectively, the study highlights the potential of quantum-enabled predictive analytics in finance and provides design and evaluation guidance for future quantum financial applications.

Abstract

Predicting loan eligibility with high accuracy remains a significant challenge in the finance sector. Accurate predictions enable financial institutions to make informed decisions, mitigate risks, and effectively adapt services to meet customer needs. However, the complexity and the high-dimensional nature of financial data have always posed significant challenges to achieving this level of precision. To overcome these issues, we propose a novel approach that employs Quantum Machine Learning (QML) for Loan Eligibility Prediction using Quantum Neural Networks (LEP-QNN). Our innovative approach achieves an accuracy of 98% in predicting loan eligibility from a single, comprehensive dataset. This performance boost is attributed to the strategic implementation of a dropout mechanism within the quantum circuit, aimed at minimizing overfitting and thereby improving the model's predictive reliability. In addition, our exploration of various optimizers leads to identifying the most efficient setup for our LEP-QNN framework, optimizing its performance. We also rigorously evaluate the resilience of LEP-QNN under different quantum noise scenarios, ensuring its robustness and dependability for quantum computing environments. This research showcases the potential of QML in financial predictions and establishes a foundational guide for advancing QML technologies, marking a step towards developing advanced, quantum-driven financial decision-making tools.

Figures (7)

• Figure 1: Overview of the LEP-QNN framework, the diagram encompasses the QNN model's interaction with various noise models, optimized by a selection of advanced optimizers within the IBM Qiskit and Pennylane platforms. This encapsulation underscores our novel contributions to QML for loan eligibility prediction.
• Figure 2: The LEP-QNN framework operational flowchart, outlining the integrated pipeline from dataset ingestion to predictive output. The initial phase encompasses a rigorous data treatment process. Central to the pipeline is the QC environment setup, where the nuances of quantum computation are discussed. The QNN Design phase is iterative, incorporating feedback from subsequent training and optimization to refine the model. The output stage is characterized by a dual focus on accuracy and interpretability, with performance metrics providing insights into the model's efficacy and MSE loss, offering a quantitative measure of prediction precision.
• Figure 3: Visualization of the QNN circuit used in the LEP-QNN framework. The full circuit is shown at the top, with each layer comprising angle encoding followed by a series of parameterized $RY$ and $RX$ gates, and entangling $CNOT$ gates arranged in a ring pattern.
• Figure 4: QNN circuit implementing dropout, with $RY$ and $RX$ gates randomly deactivated at various layers to prevent overfitting. Gates subject to dropout are marked with blue crosses (①, and ②), demonstrating the model's reduced sensitivity to specific data features through regularization.
• Figure 5: The experimental setup in our evaluation.