CONSTRUCTION AND STUDY OF A MODEL FOR ASSESSING HUMAN HEART HEALTH BASED ON THE K-NEAREST NEIGHBOR METHOD
Authors:
1.
KNITU-KAI
2.
Kazan National Research Technical University named after A.N. Tupolev
GBU «Bezopasnost' dorozhnogo dvizheniya»
GBU «Bezopasnost' dorozhnogo dvizheniya»
Type:
Article
Pages:
from 111
to 117
Status:
Published
Received:
28.01.2026
Accepted:
28.01.2026
Published:
28.01.2026
Language:
Russian
Keywords:
INTELLIGENT MODEL, CLASSIFICATION, HEART HEALTH ASSESSMENT, CARDIOVASCULAR DISEASES, K-NEAREST NEIGHBORS METHOD, MACHINE LEARNING
Abstract:
This article explores the development and study of an intelligent model for assessing human cardiac health. It is noted that diagnosing the human cardiovascular system is a pressing topic in the field of machine learning and artificial intelligence. Using machine learning algorithms to solve this problem allows for the identification of disease precursors, risk assessment, and the provision of personalized treatment and lifestyle recommendations. The concept of a healthy heart is examined, a classification of cardiovascular diseases is presented, diagnostic methods and stages are analyzed, and the feasibility of developing an intelligent model for assessing human cardiac health is discussed. The k-nearest neighbors method, used for solving classification problems, was selected for building the model. The "Heart Disease Dataset (Comprehensive)" dataset from the publicly available Kaggle resource was selected for training and testing the model. The dataset consists of 1,190 records with information on patients with cardiovascular diseases. There are two output classes: "0" (healthy) and "1" (risk of disease). The class ratio in the dataset is 47.14% ("0") and 52.86% ("1"). A correlation analysis was performed to assess the interdependence of the data. All data was randomly divided into training and test sets in an 80/20 ratio. The training set contained 952 records, and the test set contained 238. The model was built using the sklearn library in Python, as well as a tool for working with Kaggle notebooks. A P100 GPU accelerator was used as an accelerator, which significantly reduced the time for building and evaluating the model. Using the classification_report function of the sklearn library, the constructed model was evaluated using various metrics on the test data set. Metric values were calculated for each class. The results of metric calculations allowed us to conclude that the constructed model is adequate. The constructed model was also evaluated on the test data set using the AUC-ROC metric. The metric value was 0.9747. Furthermore, the accuracy of the constructed model was compared with that of other classification methods. The most popular solutions presented on the Kaggle platform were selected for comparison. The constructed model demonstrated superior results compared to other well-known methods, indicating its effectiveness. Therefore, the constructed model can be effectively used as a tool for assessing human heart health, for example, through a desktop or web application.
This article explores the development and study of an intelligent model for assessing human cardiac health. It is noted that diagnosing the human cardiovascular system is a pressing topic in the field of machine learning and artificial intelligence. Using machine learning algorithms to solve this problem allows for the identification of disease precursors, risk assessment, and the provision of personalized treatment and lifestyle recommendations. The concept of a healthy heart is examined, a classification of cardiovascular diseases is presented, diagnostic methods and stages are analyzed, and the feasibility of developing an intelligent model for assessing human cardiac health is discussed. The k-nearest neighbors method, used for solving classification problems, was selected for building the model. The "Heart Disease Dataset (Comprehensive)" dataset from the publicly available Kaggle resource was selected for training and testing the model. The dataset consists of 1,190 records with information on patients with cardiovascular diseases. There are two output classes: "0" (healthy) and "1" (risk of disease). The class ratio in the dataset is 47.14% ("0") and 52.86% ("1"). A correlation analysis was performed to assess the interdependence of the data. All data was randomly divided into training and test sets in an 80/20 ratio. The training set contained 952 records, and the test set contained 238. The model was built using the sklearn library in Python, as well as a tool for working with Kaggle notebooks. A P100 GPU accelerator was used as an accelerator, which significantly reduced the time for building and evaluating the model. Using the classification_report function of the sklearn library, the constructed model was evaluated using various metrics on the test data set. Metric values were calculated for each class. The results of metric calculations allowed us to conclude that the constructed model is adequate. The constructed model was also evaluated on the test data set using the AUC-ROC metric. The metric value was 0.9747. Furthermore, the accuracy of the constructed model was compared with that of other classification methods. The most popular solutions presented on the Kaggle platform were selected for comparison. The constructed model demonstrated superior results compared to other well-known methods, indicating its effectiveness. Therefore, the constructed model can be effectively used as a tool for assessing human heart health, for example, through a desktop or web application.
Keywords:
INTELLIGENT MODEL, CLASSIFICATION, HEART HEALTH ASSESSMENT, CARDIOVASCULAR DISEASES, K-NEAREST NEIGHBORS METHOD, MACHINE LEARNING
INTELLIGENT MODEL, CLASSIFICATION, HEART HEALTH ASSESSMENT, CARDIOVASCULAR DISEASES, K-NEAREST NEIGHBORS METHOD, MACHINE LEARNING
