Kwara State University
Kazeem Dauda
Designation: Lecturer I
Department: Statistics and Mathematical Sciences
| My Publications | ||||||||
|---|---|---|---|---|---|---|---|---|
| S/N | Title | Abstract | Authors | Volume Numbers | Publication Type | Publication Date | Link | |
| 1 | DEVELOPMENT OF HYBRIDIZED HIGH-DIMENSIONAL TECHNIQUES FOR SELECTING HIGHLY INFLUENTIAL GENES IN THE HEART RHYTHM |
An innovative hybridized model based on Genetic Algorithms-Neural Network (GANN), Borutal Algorithm-Neural Network (BANN), Genetic Algorithms, Deep Neural Network (GADNN), and Borutal Algorithm-Deep Neural Network (BADNN) have been proposed with the aim of extracting highly and differentially expressed genes; and to add an innovative knowledge to the growing domain of hybrid deep neural network and genetic programming application in the rem of heart disease pathology. The hybridized models can efficiently extract genes that are highly informative and influential from the high-dimensional dataset and have the capability of reducing the possible gene variability error by the preprocessing techniques embedded in the genetic algorithm (GA) and boruta algorithm (BA). The major contribution and novelty of this study come from three main focuses. Initially, the first and second steps emphasize feature extraction through filtering and wrapper of high-dimensional and complex data set and thirdly implement the embedded method Deep Learning. The purpose of these implementations is not only to improve classification results but also to select the most influential genes and without trading off with the most important genes. | Kazeem A. Dauda, Kabir O. Olorede, and Samuel O. Adewale | Technical Report Submitted to TETFUND Institutional Based Research, through Kwara State University, Malete. | Technical Report | 2021-06-01 | ||