1 October 2024 in Bulletin of Electrical Engineering and Informatics
Activation functions (AFs) play a critical role in artificial neural networks, allowing for the modeling of complex, non-linear relationships in data. In this review paper, we provide an overview of the most commonly used AFs in deep learning. In this comparative study, we survey and compare the different AFs in deep learning and artificial neural networks. Our aim is to provide insights into the strengths and weaknesses of each AF and to provide guidance on the appropriate selection of AFs for different types of problems. We evaluate the most commonly used AFs, including sigmoid, tanh, rectified linear units (ReLUs) and its variants, exponential linear unit (ELU), and SoftMax. For each activation category, we discuss its properties, mathematical formulation (MF), and the benefits and drawbacks in terms of its ability to model complex, non-linear relationships in data. In conclusion, this comparative study provides a comprehensive overview of the properties and performance of different AFs, and serves as a valuable resource for researchers and practitioners in deep learning and artificial neural networks.
1 January 2019 in Advances in Science, Technology and Engineering Systems Journal
30 October 2019 in International Journal of Innovative Technology and Exploring Engineering
Transformation presents the second step in the ETL process that is responsible for extracting, transforming and loading data into a data warehouse. The role of transformation is to set up several operations to clean, to format and to unify types and data coming from multiple and different data sources. The goal is to get data to conform to the schema of the data warehouse to avoid any ambiguity problems during the data storage and analytical operations. Transforming data coming from structured, semi-structured and unstructured data sources need two levels of treatments: the first one is transformation schema to schema to get a unified schema for all selected data sources and the second treatment is transformation data to data to unify all types and data gathered. To ensure the setting up of these steps we propose in this paper a process switch from one database schema to another as a part of transformation schema to schema, and a meta-model based on MDA approach to describe the main operations of transformation data to data. The results of our transformations propose a data loading in one of the four schemas of NoSQL to best meet the constraints and requirements of Big Data.