Akademik Veri Yönetim Sistemi
1st International Conference on Computing and Machine Intelligence, İstanbul, Türkiye, 19 - 20 Şubat 2021, ss.89-92
Since its emergence in 2019, the coronavirus
epidemic, which has affected the whole world, especially
Wuhan, China, continues to spread without being prevented.
Early diagnosis plays a major role in preventing and reducing
the coronavirus epidemic day by day. Covid-19 disease diagnosis
is based on many diagnostic methods such as computed
tomography, ultrasound imaging, laboratory tests. Artificial
intelligence appears as a helpful tool for these diagnostic
methods. Artificial intelligence saves time, cost and labor in
diagnosis. Today, PCR (Polymerase Chain Reaction) test is
actively used in the diagnosis of Covid-19In this study,
biochemistry parameters were used in the diagnosis of Covid-19
disease and an application was developed to determine the
priorities of biochemistry parameters in diagnosis. In the study,
feature selection process was performed with the relieff method
among all the biochemistry parameters and 6 priority
parameters were determined. The classification process was
performed with the priority parameters and 89.3% accuracy,
93.4% specificity, 85% sensitivity, 92.7% sensitivity and 88.7%
F1 score values were obtained with support vector machines. As
a result of the study, it has been observed that the classification
made with the features selected with the Relieff algorithm is
more successful than the classification made using all
biochemistry parameter parameters. It is thought that the work
carried out will help early diagnosis in the Covid-19 outbreak,
as well as reduce the workload of healthcare workers and save
costs with the help of artificial intelligence.