Title:
Adaptive Deep Learning Model for Data Normalization in Multivariate Time-Series Data
Author:
Supervisor:
Department:
Computing
Year:
2026
URL:
Abstract:
Normalization techniques are important for developing Deep Neural Network (DNN) models using time-series data because they help make training more stable and efficient by adjusting the features to sim...
View more
Multimedia Content:
01
Methodology
1. PRISMA-Guided Literature Review
A comprehensive literature review was conducted to investigate existing normalization techniques, deep adaptive learning methods, time-series forecasting approach...
View more
02
Citation Analysis
1. IEEE Transactions on Neural Networks and Learning Systems
Scope: Neural networks, deep learning, adaptive learning systems, and machine learning models.
Relevance: Highly relevant because the t...
View more
03
Impact
1. Technical Impact
This research introduces the Adaptive Centering Normalization (ACN) framework, a novel adaptive normalization layer specifically designed to address the challenges of non-statio...
View more
04
Benefits to Society
1. Environmental Benefits
The proposed forecasting framework can enhance predictive capabilities for environmental and metocean applications by improving the analysis of oceanographic and weather-r...
View more
05
Applications
1. Financial Market Forecasting
Benefit: Improves forecasting accuracy for complex financial time-series data.
Implementation: The ACN framework can be integrated into trading systems, investment ...
View more
06
Infographic
Summary of this Thesis
07
Patents / Products
Product Name:
Adaptive Centering Normalization (ACN) Framework
Significance:
The ACN framework represents the main innovation of the thesis by introducing a dynamic normalization mechanism capabl...
View more
08
Experts
1. AP Ts Dr Hitham Seddiq A. A.
Institution: Universiti Teknologi PETRONAS (UTP), Malaysia
Contribution: Main supervisor for the development of the Adaptive Centering Normalization framework and d...
View more