Uma especificação do algoritmo genético baseado em tipos abstratos de dados para diagnóstico de arritmias ventriculares a ser aplicada a partir do eletrocardiograma

Abstract

Diagnosis Computer Systems are softwares used to aid medical professionals to formulate an opinion about how a specific human organ works. Therefore, these systems receive as input parameters of normality obtained from literature, which will be confronted with the history and / or exams of patients, generating a diagnosis with the presence or absence of infirmity. The data and exams of patients can be randomly generated, extracted by means of signal or image from reliable data banks, such as MIT, or obtained by the exams equipment itself (for example: electrocardiograph), when there is a module with a USB port, or simply by inserting numbers arranged in a Cartesian plane. This work presents a computational system for diagnosing Ventricular Arrhythmias which use an evolutionary computation approach, more specifically the genetic algorithm based on abstract data types, which aims to detect abnormalities present on exam of the electrocardiogram (ECG) by a careful analysis of information provided to System Electrocardiogram (SIECG). The proposed system covers the events of the ECG (waves, complexes, intervals and segments) with their respective peripheral leads and precordial, besides profile of the patient, which comprises the set of your personal data ( age, sex). The SIECG has no USB port, since it aims to meet the regional public health, which has electrocardiographs old, so it will be powered by figures from the ECG, which correspond to the height and length of electrocardiographic events. The tests generated for the issuance of the report of the medical team, which will be compared with the proposed system was intentionally designed with the presence of three ventricular arrhythmias, namely: Right Bundle Branch Block (BBBR), Left Bundle Branch Block (BBBL) and ventricular extrasystole. This system was developed in Python and Mysql.