Autism Spectrum Disorder is a mental disorder that afflicts millions of people worldwide. It is estimated that one in 160 children has traces of autism, with five times the higher prevalence in boys. The protocols for detecting symptoms are diverse. However, the following are among the most used: the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5), of the American Psychiatric Association; the Revised Autistic Diagnostic Observation Schedule (ADOS-R); the Autistic Diagnostic Interview (ADI); and the International Classification of Diseases, 10th edition (ICD-10), published by the World Health Organization (WHO) and adopted in Brazil by the Unified Health System (SUS). The application of machine learning models helps make the diagnostic process of Autism Spectrum Disorder more precise, reducing, in many cases, the number of criteria necessary for evaluation, denoting a form of attribute engineering (feature engineering) efficiency. This work proposes a hybrid approach based on machine learning algorithms’ composition to discover knowledge and concepts associated with the multicriteria method of decision support based on Verbal Decision Analysis to refine the results. Therefore, the study has the general objective of evaluating how the mentioned hybrid methodology proposal can make the protocol derived from ICD-10 more efficient, providing agility to diagnosing Autism Spectrum Disorder by observing a minor symptom. The study database covers thousands of cases of people who, once diagnosed, obtained government assistance in Brazil.

Predictive modelling strategies can optimise the clinical diagnostic process by identifying patterns among various symptoms and risk factors, such as those presented in cases of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), also known as coronavirus (COVID-19). In this context, the present research proposes a comparative analysis using benchmarking techniques to evaluate and validate the performance of some classification algorithms applied to the same dataset, which contains information collected from patients diagnosed with COVID-19, registered in the Influenza Epidemiological Surveillance System (SIVEP). With this approach, 30,000 cases were analysed during the training and testing phase of the prediction models. This work proposes a comparative approach of machine learning algorithms (ML), working on the knowledge discovery task to predict clinical evolution in patients diagnosed with COVID-19. Our experiments show, through appropriate metrics, that the clinical evolution classification process of patients diagnosed with COVID-19 using the Multilayer Perceptron algorithm performs well against other ML algorithms. Its use has significant consequences for vital prognosis and agility in measures used in the first consultations in hospitals.

Artificial Intelligence techniques based on Machine Learning algorithms, Neural Networks and Naïve Bayes can optimise the diagnostic process of the SARS-CoV-2 or Covid-19. The most significant help of these techniques is analysing data recorded by health professionals when treating patients with this disease. Health professionals' more specific focus is due to the reduction in the number of observable signs and symptoms, ranging from an acute respiratory condition to severe pneumonia, showing an efficient form of attribute engineering. It is important to note that the clinical diagnosis can vary from asymptomatic to extremely harsh conditions. About 80% of patients with Covid-19 may be asymptomatic or have few symptoms. Approximately 20% of the detected cases require hospital care because they have difficulty breathing, of which about 5% may require ventilatory support in the Intensive Care Unit. Also, the present study proposes a hybrid approach model, structured in the composition of Artificial Intelligence techniques, using Machine Learning algorithms, associated with multicriteria methods of decision support based on the Verbal Decision Analysis methodology, aiming at the discovery of knowledge, as well as exploring the predictive power of specific data in this study, to optimise the diagnostic models of Covid-19. Thus, the model will provide greater accuracy to the diagnosis sought through clinical observation.

No abstract