نوع مقاله : مقاله پژوهشی
نویسنده
دانشکده مهندسی کامپیوتر - دانشگاه آزاد اسلامی واحد میبد - میبد - ایران
چکیده
پیشینه و اهداف: بیماریهای قلبی بهعنوان یکی از شایعترین بیماریهای جهان معرفیشدهاند و تأخیر در درمان آنها میتواند منجر به افزایش مرگ بیماران گردد. هدف اصلی این تحقیق ارتقاء شناسایی بیماران قلبی با استفاده از سیستم دستهبند یادگیر است.
روشها: در این تحقیق، از سیستمهای دستهبند یادگیر با تکنیکهای یادگیری مبتنی بر قواعد استفادهشده است. این تکنیکها بر پایه دو اصل اساسی یادگیری تقویتی و الگوریتمهای تکاملی ژنتیک ساختهشدهاند. سبک میشیگان بهعنوان روش بهینهسازی انتخابشده و مجموعه داده بیماران قلبی از مرکز تحقیقات افشار برای آموزش و یادگیری سیستم مورداستفاده قرارگرفته است.
یافتهها: پس از آموزش سیستم، تعدادی قانون باارزش تولیدشده که در مرحله آزمون برای پیشبینی بیماران قلبی مورداستفاده قرار گرفتهاست. نتایج آزمایشها نشانمیدهد که با استفاده از سیستم دستهبند یادگیر بر مبنای سبک میشیگان، شناسایی بیماران قلبی بهبودیافته و دقت پیشبینی به ۸۸ درصد افزایشیافته است؛ که این روش قادر به انجام شناسایی کاملتری از بیماران قلبی است.
نتیجهگیری: با توجه به نتایج تحقیق، استفاده از سیستم دستهبند یادگیر بر مبنای سبک میشیگان بهعنوان یک رویکرد بهینه، شناسایی بیماران قلبی را بهبود بخشیده و دقت پیشبینی را افزایش داده است. این روش میتواند بهبود مؤثری در درمان بهموقع بیماران قلبی و کاهش مرگومیر ناشی از این بیماریها داشته باشد.
کلیدواژهها
عنوان مقاله [English]
Evaluating the Performance of a Machine Learning Classifier System for the Identification of Heart Disease Patients
نویسنده [English]
- mohammadreza dehghani mahmoudabadi
Faculty of Computer Engineering, University of Azad Islamic Maybod, Maybod, Iran,
چکیده [English]
Background and Objectives: Cardiovascular diseases have been identified as one of the most prevalent global health issues, and delays in treatment can lead to increased mortality among patients. The primary objective of this study has been to enhance the identification of heart disease patients using a machine learning classification system.
Methods: In this research, machine learning classification systems with rule-based learning techniques have been employed. These techniques are built upon two fundamental principles, reinforcement learning, and genetic algorithms. The Mishgan style has been selected as the optimization method, and a dataset of heart disease patients from the Afshar Research Center has been utilized for the training and learning of the system.
Findings: Following the training of the system, a set of valuable rules has been generated and utilized in the testing phase for predicting heart disease patients. The experimental results indicate that using the Mishgan-style machine learning classification system has improved the identification of heart disease patients, resulting in an 88% increase in prediction accuracy. In other words, this approach enables a more comprehensive identification of heart disease patients.
Conclusion Considering the study's outcomes, the use of the Mishgan-style machine learning classification system as an optimal approach has enhanced the identification of heart disease patients and increased prediction accuracy. This method can contribute significantly to timely treatment of heart disease patients and the reduction of morbidity and mortality associated with these diseases.
کلیدواژهها [English]
- Cardiovascular Diseases
- Learning Classifier System
- Rule-Based Learning