Loading...

Heart Disease Diagnosis Based on Heart Sounds Using Signal Processing and Machine Learning Algorithms

Zeinali, Yasser | 2020

326 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 53275 (01)
  4. University: Sharif University of Technology
  5. Department: Industrial Engineering
  6. Advisor(s): Akhavan Niaki, Taghi
  7. Abstract:
  8. The research in this study aims to analyze data in healthcare, especially the diagnosis of several diseases caused by heart failure. Analyzing and analyzing this data can lead to the discovery of relationships and patterns that can play an important role in the decision-making process of relevant officials in any field. Today, medical data around the world is stored in large volumes for future research. Various infrastructures and software have been set up in many health centers and research centers affiliated with those organizations.In this research, the general process of work is such that the data related to the heart sounds, which are in the four broad categories of S1 to S4, are identified and classified by the heart sounds. The sounds S1 and S2 are the natural sounds of the heart, and the sounds S3 and S4 are the abnormal sounds of the heart, each of which expresses a particular type of heart disease. In the first step, after retrieving the data by signal processing algorithms, we extracted their properties. We then compiled a set of data contacting the extracted features for the decision-making process. In the next step, we used the feature selection algorithms to select the effective features so that we can get the optimal answer faster in addition to reducing the dimensions of the problem. In the final section, some of the most popular classification algorithms were used to classify the type of sound.The innovation of this research is to classify sounds so that in addition to determining whether they are normal or abnormal, it will be apparent that if the sound is abnormal, it has the sound of S3 or S4. One of the most significant obstacles is that we do not know the label of each category. So to solve this problem, we acted in two ways. One method was to use the opinion of a specialist, and the other was to use clustering algorithms, which are described in detail in the following sections. Finally, we achieved 87.5% accuracy in classifying the sounds of 3 classes and improved the results of previous research by 6%
  9. Keywords:
  10. Heart Sound ; Classification ; Clustering ; Machine Learning ; Feature Selection ; Feature Extraction ; Signal Processing ; Heart Diseases

 Digital Object List

 Bookmark

No TOC