Yazar "Alkhafaji, Mohammed Jasim A." seçeneğine göre listele

Listeleniyor 1 - 1 / 1
  • [ X ]
    Öğe
    Clean medical data and predict heart disease
    (Institute of Electrical and Electronics Engineers Inc., 2020) Alkhafaji, Mohammed Jasim A.; Aljuboori, Abbas Fadhil; Ibrahim, Abdullahi Abdu
    The enormous data provided by the health care environment needs many important and powerful tools for analyzing and extracting data and accessing useful knowledge. Many researchers have been interested in applying many statistical tools as well as many different data mining tools in order to improve an analysis process and extract data from a different data set. The only thing that proves the success and robustness of data mining tool is accurate diagnosis of the disease. According to the (WHO), the biggest cause of death in the last ten years or so in this vast world is heart disease. The statistical exploration tools that researchers use are tools that help decision-makers in health care to predict and diagnose heart disease. The tools used in the diagnostic process for heart disease have been thoroughly tested in order to demonstrate sufficient and acceptable accuracy. A set of patient data divided into 665 records was used, of which 300 were for males, with 365 for females, with 10 different related characteristics. The decision-making department still suffers from a lack of performance and decision-making. Our paper aims to process data in different ways before the process of accessing knowledge to make the appropriate decision through expectations of classification analysis and then using techniques to extract data with acceptable accuracy. Our goal proposed in this paper is to purify the data before the disease prediction process to get the best possible prediction and compare the results with the results of a group of previous researchers to reach an accurate diagnosis and prediction. The second part of our goal is to compare between different technologies on different data sets such as decision tree technology and the second technique is Bayesian classification technology and the last technology is neural networks and the results were (98.85%, 98.16%, 91.31%), respectively. In the end, we hope to obtain acceptable results with high accuracy in the future, enhance clinical diagnosis, and promote appropriate decision-making for early treatment specialists. © 2020 IEEE.