Hershey, Pennsylvania
New York, New YorkBeijing, China
Optimization Techniques for Influenza Prediction in Biological Expert Systems

Optimization Techniques for Influenza Prediction in Biological Expert Systems

U. Vignesh (Vellore Institute of Technology, Chennai, India) and Rahul Ratnakumar (Manipal Institute of Technology, India)
Copyright: © 2024 |Pages: 15
DOI: 10.4018/979-8-3693-1131-8.ch010
OnDemand:
(Individual Chapters)
Available
$37.50
Current Special Offers
No Current Special Offers
TOTAL SAVINGS: $37.50

Abstract

Currently, the biggest challenge in the world is the detection of viral infection in various diseases, as par to the rapid spread of the disease. According to recent statistics, the number of people diagnosed with the Influenza virus is exponentially increasing day by day, with more than 2.5 million confirmed cases. The model proposed here analyses the Influenza virus by comparing different deep learning algorithms to bring out the best in terms of accuracy for detection and prediction. The models are trained using CT scan dataset comprising of both Influenza positive patients and negative patients. The results of algorithms are compared based on parameters such as train accuracy, test loss, etc. Some of the best models after training were, DenseNet-121 with accuracy of 96.28%, VGG-16 with accuracy of 95.75%, ResNet-50 with accuracy of 94.18%, etc. in detecting the virus from the CT scan dataset with the proposed ACDL algorithm. Thus, these models will be helpful and useful to the government and communities to initiate proper measures to control the outbreak of the Influenza virus in time.
Chapter Preview
Top

Wang et al. (2020) used transfer learning and model integration based on the COVID-net model by Darwin-AL for identifying the COVID virus from Chest X-RAY images, the methodologies were implemented using ResNet-101 and ResNet-152 neural networks. Rajaraman et al. (2020) used iterative pruning for identifying pulmonary signs of the virus, for this they came up with a CNN and set of pre-trained models that have undergone training and evaluation based on Chest X-rays to learn about feature representations concerning modality. Hassan et al. (2020) came up with COVID-19 prediction system using deep learning on chest X-ray images. Mohammed et al. (2020) came up with a method for organizations in selecting a COVID-19 diagnosis system based on the multi-criteria decision-making (MCDM) method. Ko et al. (2020) used a 2D framework known as the FCO-net model, which is a technique to diagnose the virus from images based on CT and to segregate it from non–COVID-19 diseases and non-pneumonia diseases. Transfer learning is used to create the FCO-Net model with the help of four models which are VGG16, ResNet-50, Inception-v3, and Xception. Alazab et al. (2020) presented an artificial-intelligence technique based on CNN to detect COVID19 patients using CXR as a dataset. Punn et al. (2020) used DNN and RNN with the help of LSTM cells for predicting the number of recovered, confirmed, and death cases worldwide. Mahalle et al. (2020) made use of multimodal data for COVID-19 prediction.

In this study, a single dataset is used to implement (train and test) the necessary deep learning algorithms. A CT scan dataset comprising of Influenza patients and healthy persons is used (746 images in total).

Complete Chapter List

Search this Book:
Reset