Open AccessFormation of descriptors for medical risk classifiers based on the current-voltage characteristics of biologically active points
Author(s) -
А. В. Мирошников,
Alexander Kiselev,
О. В. Шаталова,
Sofiya Kadyrova
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/2060/1/012013
Subject(s) - artificial neural network , classifier (uml) , computer science , artificial intelligence , current (fluid) , machine learning , electrical current , data mining , pattern recognition (psychology) , engineering , electrical engineering
Diagnosis of many diseases can be achieved using the electrical response of bioactive points on patient’s body. Predictions are made by analyzing the electrical response of patients’ biologically active points. To analyze the impact on the organ, values of the current strength in the organ were measured and the changes of the passive electrical properties on selected bioactive locations on the body. The research goal is to develop a classification neural network model to predict post-surgery complication risk for prostate surgery. The obtained current-voltage characteristics is used as input in the classification artificial neural network model that predicts two classes. The dataset involves medical information of 120 patients diagnosed with prostate disease. Descriptors are composed of the analysis of five current-voltage characteristics. Neural network classifiers prediction shows a high capacity to predict critical conditions. The developed model is valuable because it can be used in beepers for prenosological diagnosis of infectious diseases. The 3 BAP models using I – Taiyuang, IX – Daling, and V – Shenmen configuration produced the highest accuracy. The accuracy of the classifier ranged from 80% to 82%.