Content area
Full text
Abstract
This study investigates the application of LSTM Neural Networks in forecasting viral spread, leveraging CDC's state-level data. It seeks to determine the most precise neural network model for predicting changes in viral strains, focusing on utilizing several different artificial intelligence platforms. After several iterations of each type of model, MSE and MAPE values were used to evaluate performance. LSTM performed better over neural networks, but Spiking Neural Networks (SNN) significantly improved over both methods. Further investigation for SNN models should yield more reliable results.
Keywords
Neural Networks, Virus Prediction, LSTM, Spiking Neural Networks, Influenza
1. Introduction
Viral infections pose significant threats to public health and the economy[4]. Accurate prediction models are crucial for effective response and resource allocation [5]. This study explores neural network-based models to forecast virus spread and mutation rates. This study builds upon the work of Hamilton et al., who focused on developing a neural network model to predict viral load utilizing spiking neural networks [1]. This study aims to compare the effectiveness of utilizing Neural Networks, Long Short-Term Memory Neural Networks, and Spiking Neural Networks with viral load prediction.
2. Background
2.1 Neural Networks
Neural networks are a fundamental component of artificial intelligence, inspired by the structure and function of the human brain. They consist of interconnected nodes or neurons, which process information in a manner akin to biological neurons. These networks are designed to recognize patterns and make decisions by learning from data[2]. The strength of neural network connections is adjusted during training, allowing the network to improve its performance on specific tasks over time.
2.2 Long Short-Term Memory Networks
Long Short-Term Memory (LSTM) networks...