Enhanced Predictive Data Modeling for Specialized Sciences using Least Squares Convex Optimization
Year: 2025
Type: article
Abstract: This paper introduces an improved approach to predictive data modeling by leveraging least squares optimization and global convex analysis. We begin with the construction of a linear predictive model and apply the least squares method to minimize residual error. Subsequently, we incorporate global convex optimization techniques to refine the model using quadratic forms. This approach offers enhanced prediction accuracy and robustness for specialized scientific datasets. The methodology is further translated into algorithmic pseudocode suitable for large-scale data programming. Real-world examples and visual illustrations validate the efficacy of the proposed model.
Author: Tran Dang Hung
Institution: Ho Chi Minh City University of Industry and Trade
Language: English
Subfield: Applied Mathematics
Field: Mathematics
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4411065910