Predicting infection with coronavirus wirelessly based on artificial neural network and MATLAB DOI Open Access

Suha Dalaf Fahad,

Sadik Kamel Gharghan,

Raghad Hassan Hussein

et al.

AIP conference proceedings, Journal Year: 2023, Volume and Issue: 2591, P. 020010 - 020010

Published: Jan. 1, 2023

Due to the huge fast spread of Covid-19 around world, which resulted in loss many lives, maximum level emergency was triggered all over world. The best way reduce COVID-19 infection is prediction it early based on artificial intelligence (AI). To determine whether patient has or not. An accurate and effective diagnosis system for proposed this paper. diagnostic parameters right left lungs, D-dimer, physiological such as SpO2, temperature, heart rate were collected from CT scans (three RGB colors) rate. data 300 patients, with each receiving 10 samples; 114 them infected Covid-19, while remaining 186 uninfected. For training, testing, verifying gathered data, an neural network (ANN) one hidden layer at 20 nodes-based Backpropagation method used. parameters, a total 30,000 samples obtained (300 x per patient). 3,000 individuals, each) divided into three datasets: 70% training ANN (2,100 out samples), 15% testing (450 validation samples) samples). In terms performance, correlation coefficient, error, mean absolute error (MAE), histogram, results studied. MAE validation, nodes, respectively, 0.0012, 0.012, 0.013, indicating that achieves good accuracy. coefficient (R2) between actual estimated value 0.9999, 0.9996, 0.9998, respectively. MAE, suggested technique beat current state-of-the-art.

Language: Английский

Skin and Body Temperature Parameter Calibration of MAX30100 Sensor Module Based on Arduino-Uno DOI Open Access

Juan Karnadi,

Arbi Riantono,

Ibnu Roihan

et al.

Evergreen, Journal Year: 2024, Volume and Issue: 11(2), P. 1419 - 1425

Published: June 1, 2024

ic of temperature change's aspect with DS18B20 sensor in skin reading. The calibration effort's process has four leading stages sequencingly: averaging, filtering, fitting (nearing) towards temperature's reading, and lastly nearing the result to body thermometer as calibrator. respectively ammount 6.02°C for 1.34°C temperature. According these following results, error rate is ±0.27% thermometer.

Language: Английский

Citations

0
Predicting infection with coronavirus wirelessly based on artificial neural network and MATLAB DOI Open Access

Suha Dalaf Fahad,

Sadik Kamel Gharghan,

Raghad Hassan Hussein

et al.

AIP conference proceedings, Journal Year: 2023, Volume and Issue: 2591, P. 020010 - 020010

Published: Jan. 1, 2023

Due to the huge fast spread of Covid-19 around world, which resulted in loss many lives, maximum level emergency was triggered all over world. The best way reduce COVID-19 infection is prediction it early based on artificial intelligence (AI). To determine whether patient has or not. An accurate and effective diagnosis system for proposed this paper. diagnostic parameters right left lungs, D-dimer, physiological such as SpO2, temperature, heart rate were collected from CT scans (three RGB colors) rate. data 300 patients, with each receiving 10 samples; 114 them infected Covid-19, while remaining 186 uninfected. For training, testing, verifying gathered data, an neural network (ANN) one hidden layer at 20 nodes-based Backpropagation method used. parameters, a total 30,000 samples obtained (300 x per patient). 3,000 individuals, each) divided into three datasets: 70% training ANN (2,100 out samples), 15% testing (450 validation samples) samples). In terms performance, correlation coefficient, error, mean absolute error (MAE), histogram, results studied. MAE validation, nodes, respectively, 0.0012, 0.012, 0.013, indicating that achieves good accuracy. coefficient (R2) between actual estimated value 0.9999, 0.9996, 0.9998, respectively. MAE, suggested technique beat current state-of-the-art.

Language: Английский

Citations

0