Risk Parity And Dynamic Portfolio Allocation- Integrated Personalized Financial Advice Provider Model Using RL-GPLinQ-GRU and IGP-Fuzzy

Abstract

Nowadays, individuals are increasingly seeking accessible and intelligent tools to help and manage their finances for better financial decision-making. However, traditional works failed to consider severe traditional assets, spending priority, real-time financial data, and market feedback to perform risk parity and dynamic allocation, leading to ineffective portfolio adjustment. Therefore, an efficient risk parity and dynamic portfolio allocation-integrated personalized financial advice provider framework is proposed using RL-GPLinQ-GRU and IGP-Fuzzy. Initially, financial data undergoes preprocessing, augmentation, bubble plot generation, and feature extraction.  Meanwhile, the MT-DBSCAN is used to analyze behavioral patterns of preprocessed financial data, followed by portfolio extraction and portfolio optimization using PS-MPA. Now, RL-GPLinQ-GRU identifies financial scores from the extracted features and optimized portfolio via enhancing model integrity using VLIME. In the meantime, the sentiment classification model is trained using the stock market sentiment data through preprocessing, keyword extraction, word embedding, and sentiment analysis. Simultaneously, RL-GPLinQ-GRU predicts stock market status using the historical stock market data via preprocessing, technical indicator extraction, and feature extraction. Finally, based on the identified financial score with explanation, classified sentiments, and predicted stock market status, IGP-Fuzzy performs risk parity and dynamic portfolio allocation. Thus, the proposed framework outperforms the other prevailing techniques by attaining higher accuracy (99.21%).