원문정보
초록
영어
There are serious problems worldwide, such as a pandemic due to an unprecedented infection caused by COVID-19. On previous approaches, they invented medical vaccines and preemptive testing tools for medical engineering. However, it is difficult to access poor medical systems and medical institutions due to disparities between countries and regions. In advanced nations, the damage was even greater due to high medical and examination costs because they did not go to the hospital. Therefore, from a software engineering-based perspective, we propose a learning model for determining coronavirus infection through symptom data-based software prediction models and tools. After a comparative analysis of various models (decision tree, Naive Bayes, KNN, multi-perceptron neural network), we decide to choose an appropriate decision tree model. Due to a lack of data, additional survey data and overseas symptom data are applied and built into the judgment model. To protect from thiswe also adapt human normalization approach with traditional Korean medicin approach. We expect to be possible to determine coronavirus, flu, allergy, and cold without medical examination and diagnosis tools through data collection and analysis by applying decision trees.
목차
1. Introduction
2. Related Works
2.1 A Prototype model for comparative prediction of COVID-19 and pandemic influenza
2.2 A Decision Tree Classifier
2.3 Naïve Bayesian
2.4 K-Nearest Neighbors (KNN) Classifier
2.5 Multilayer Perceptron on Neural Network Model
2.6 Comparison between Diverse Models with Machine Learning Algorithms.
2.7 Traditional Clinical Mechanism Research on Human Type Classification
3. Comparative Prediction Prototype Model
3.1 Data Collection and Design
3.2 Data Preprocessing
3.3 Structuring the Model
3.4 Model Visualization
4. Bio-current pattern-based prevention guide for pulmonary infectious diseases
5. Conclusion
Acknowledgement
References