Deep learning algorithms for enhancing securities price prediction and insurance strategy optimization
The process of attempting to estimate the future prices of particular stocks by utilizing historical data and various analytical tools, including deep learning algorithms, is called stock price prediction. Insurance providers’ overall approach and decisions to manage their risks, enhance their profi...
Gespeichert in:
Veröffentlicht in: | Journal of intelligent & fuzzy systems 2023-12, Vol.45 (6), p.10369-10379 |
---|---|
1. Verfasser: | |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | The process of attempting to estimate the future prices of particular stocks by utilizing historical data and various analytical tools, including deep learning algorithms, is called stock price prediction. Insurance providers’ overall approach and decisions to manage their risks, enhance their profitability, and give value to their policyholders are referred to as the insurance strategy. It requires various things to be considered, including underwriting procedures, pricing strategies, product creation, risk analysis, claims administration, and investment choices. This study proposed optimizing an insurance strategy and predicting securities prices using a deep learning algorithm. Initially, the real stock data sources for Microsoft Corporation (MSFT) were gathered from Ping An Insurance Company of China (PAICC) and the Shanghai-based National Association of Securities Dealers Automated Quotation (NASDAQ). Normalization is the procedure used to preprocess data for the raw data. We suggest an Enhanced dragonfly-optimized deep neural network (EDODNN) with stock price forecasting and insurance. The outcomes demonstrate that the proposed model outperforms the current methodology and achieves accuracy, precision, recall, F1 score, R2, and RMSE. To display the effectiveness of the suggested system, its performance is compared to more established methods to obtain the highest level of efficiency for the research. |
---|---|
ISSN: | 1064-1246 1875-8967 |
DOI: | 10.3233/JIFS-234292 |