@article{Kaewlek_Sakaekhum_Promton_Tharama_Chusin_Yabsantia_Udee_2024, title={Prediction of COVID-19 with Statistical Data on Chest Radiography using Artificial Intelligence}, volume={24}, url={https://asianmedjam.com/index.php/amjam/article/view/1423}, abstractNote={<p><strong>Introduction:</strong> COVID-19 is rapidly spreading around the world and has a high mortality rate. Artificial intelligence (AI) technology is a method that can be used to diagnose the presence of COVID-19 via chest radiographic apparatus. AI can be found to provide accurate results and increased diagnostic efficiency.<br /><strong>Objectives:</strong> To evaluate the efficacy of artificial intelligence for COVID-19 diagnosis using statistical data from radiographic chest images.<br /><strong>Methods:</strong> The research population sample consisted of 10,000 normal heathy individuals and 10,000 COVID-19 chest radiographs of patients were used for training (70.0%), validating (20.0%), and testing (10.0%). The images were segmented into the left and right lung regions by using the U-net architecture and then statistical data was calculated, including integrated density, mean, standard deviation, skewness, and kurtosis. Three artificial intelligence methods (support vector machine, K-mean clustering, and restricted Boltzmann machine) were compared the models’ predictions. The performance of three methods were analyzed for accuracy, sensitivity, specificity, precision, and F1-score.<br /><strong>Results:</strong> The accuracy of the support vector machine, K-mean clustering, and restricted Boltzmann machine were 70.5%, 62.5%, and 63.2%, respectively. The trend of the sensitivity, specificity, precision, and F1-score were similar in terms of accuracy, sensitivity, specificity, precision, and F1-score of the support vector machine, which were 64.2%, 73.5%, 68.2%, and 68.5%, respectively.<br /><strong>Conclusions:</strong> The most successful technique for diagnosing COVID-19 from chest radiographs was the support vector machine. It outperformed the restricted Boltzmann machine, which was followed by K-mean clustering</p>}, number={1}, journal={Asian Medical Journal and Alternative Medicine}, author={Kaewlek, Titipong and Sakaekhum, Waritsara and Promton, Warisa and Tharama, Areeya and Chusin, Thunyarat and Yabsantia, Sumalee and Udee, Nuntawat}, year={2024}, month={Apr.}, pages={39–48} }