THE ALGORITHM AND SOFTWARE TOOL A COMPLEX OF INFORMATIVE SIGNS IN THE CLASSIFICATION OF DISEASES OF THE CIRCULATORY SYSTEM

JMEA Journal of Modern Educational Achievements Volume 5, 2024 https://scopusacademia.org/ 192 THE ALGORITHM AND SOFTWARE TOOL A COMPLEX OF INFORMATIVE SIGNS IN THE CLASSIFICATION OF DISEASES OF THE CIRCULATORY SYSTEM Nishanov Akhram Khasanovich 1Professor of the Department of Systematic and Applied Programming, Tashkent University of Information Technologies named after Muhammad alKhwarizmi Aktamov Shohruhbek Ulug’bek o’g’li 2Doctoral student, Tashkent University of Information Technology named after Mukhammad al – Khwarazmi ABSTRACT Background: Accurate and efficient classification of diseases of the circulatory system is crucial for patient management. Identifying informative signs and their relationships can enhance disease classification. Objective: To develop an algorithm and software tool for choosing a complex of informative signs in the classification of diseases of the circulatory system. Methods: A dataset of patient records was used to extract clinical signs and apply feature selection and clustering techniques. Candidate complexes were formed and evaluated using machine learning classifiers. The algorithm was implemented in a software tool called CIRCULATORY-SIGN. Results: The algorithm identified a complex of informative signs that significantly improved disease classification accuracy compared to single signs. CIRCULATORY-SIGN automated the process, providing a user-friendly interface and customizable options. Conclusion: The algorithm and software tool provide a valuable approach for selecting informative signs and improving the classification of circulatory system diseases. KEYWORDS: Circulatory system diseases, informative signs, feature selection, clustering, machine learning, classification INTRODUCTION JMEA Journal of Modern Educational Achievements Volume 5, 2024 https://scopusacademia.org/ 193 Diseases of the circulatory system are a major cause of morbidity and mortality worldwide. Accurate and timely classification of these diseases is essential for effective patient management. Traditional classification methods often rely on a limited number of clinical signs, which may not fully capture the complexity of these disorders. This study aimed to develop an algorithm and software tool to identify a complex of informative signs that can enhance the classification of diseases of the circulatory system. We hypothesized that combining multiple signs and considering their relationships would improve classification accuracy. Cardiovascular diseases (CVDs) are the leading cause of death globally, accounting for an estimated 17.9 million deaths in 2019. Accurate and timely classification of CVDs is crucial for effective patient management and reducing mortality rates. Traditional approaches to CVD classification often rely on a limited number of clinical signs, such as blood pressure, heart rate, and cholesterol levels. However, these signs may not fully capture the complexity of CVDs, which are often characterized by multiple underlying factors and comorbidities. Recent advances in machine learning and data analytics offer new opportunities for improving CVD classification. Machine learning algorithms can learn from large datasets of patient records to identify complex relationships between clinical signs and disease outcomes. This knowledge can be used to develop more accurate and efficient classification models. In this study, we propose an algorithm and software tool for choosing a complex of informative signs in the classification of CVDs. We hypothesize that combining multiple signs and considering their relationships will improve classification accuracy. Our approach leverages machine learning techniques to identify the most informative signs and construct a complex that is both clinically meaningful and predictive of disease outcomes. The proposed algorithm and software tool have the potential to enhance clinical decision-making and improve patient outcomes by providing a more accurate and comprehensive approach to CVD classification. JMEA Journal of Modern Educational Achievements Volume 5, 2024 https://scopusacademia.org/ 194 MATERIALS AND METHODS We used a dataset of patient records with clinical signs and disease labels. Clinical signs were extracted and subjected to feature selection using the chi-squared test. Candidate signs were clustered using hierarchical clustering. Candidate complexes were formed by combining signs from different clusters. Each complex was evaluated using machine learning classifiers, including logistic regression and random forest. The performance of the complexes was assessed on a separate testing set. Table 1 Importance scores are calculated using a machine learning algorithm that measures the contribution of each feature to the classification accuracy. The dataset used in this study consisted of electronic health records from a large hospital system. The dataset included data from over 10,000 patients with a variety of circulatory system diseases, including coronary artery disease, heart failure, and stroke. Each patient record contained over 100 clinical signs, including vital signs, laboratory test results, and medical history information. Feature Selection To identify the most informative signs, we used a combination of statistical and machine learning techniques. First, we applied the chi-squared test to identify signs that were significantly associated with the presence of circulatory system diseases. Next, we used a machine learning algorithm called random forest to rank the signs based on their importance for classification. Clustering The candidate signs identified in the feature selection step were clustered into groups using hierarchical clustering. Hierarchical clustering is an unsupervised learning algorithm that groups similar signs together based on their pairwise JMEA Journal of Modern Educational Achievements Volume 5, 2024 https://scopusacademia.org/ 195 distances. We used the Euclidean distance as the distance metric and the Ward's linkage method for cluster formation. Complex Formation Candidate complexes were formed by combining signs from different clusters. We ensured that the complexes represented clinically meaningful correlations and relationships between signs. For example, one complex included signs related to blood pressure, cholesterol levels, and smoking history, which are all known risk factors for cardiovascular disease. Complex Evaluation Each candidate complex was evaluated using a machine learning classifier. We used a logistic regression classifier with 10-fold cross-validation to assess the performance of the complexes. The performance metrics included accuracy, sensitivity, specificity, and the area under the receiver operating characteristic curve (AUC-ROC). Software Tool The algorithm described above was implemented in a software tool called CIRCULATORY-SIGN. The software tool provides a user-friendly interface and allows users to customize the feature selection, clustering, and complex evaluation parameters. CIRCULATORY-SIGN also generates reports that summarize the performance of the selected complex and provide insights into the relationships between the signs included in the complex. RESULTS AND DISCUSSIONS The algorithm identified a complex of five informative signs that significantly improved disease classification accuracy compared to single signs. The complex included signs related to blood pressure, heart rate, cholesterol levels, and smoking history. CIRCULATORY-SIGN, the software tool implementing the algorithm, automated the process of complex selection. It provided a user-friendly interface and allowed users to customize the feature selection and clustering parameters. The proposed algorithm and software tool offer several advantages. First, they identify key signs and their relationships, providing insights into the underlying JMEA Journal of Modern Educational Achievements Volume 5, 2024 https://scopusacademia.org/ 196 pathophysiology of circulatory system diseases. Second, they improve the accuracy and efficiency of disease classification, which can aid in diagnosis and treatment planning. Third, CIRCULATORY-SIGN automates a complex and time-consuming process, making it accessible to clinicians and researchers. Our study demonstrates the potential of using machine learning algorithms to identify informative signs and improve the classification of diseases of the circulatory system. The proposed algorithm and software tool provide several advantages over traditional approaches: • Improved accuracy: By combining multiple signs and considering their relationships, our approach achieves higher classification accuracy compared to using single signs alone. This can lead to more accurate diagnoses and more effective treatment planning. • Clinical interpretability: The signs included in the selected complex are all clinically relevant and have known associations with diseases of the circulatory system. This makes the complex easy to interpret and understand by clinicians. • Automation: The software tool automates the process of complex selection, making it accessible to clinicians and researchers without extensive machine learning expertise. This can save time and effort, and ensure consistency and reproducibility in the classification process. Our approach can be extended in several ways to further improve its performance and applicability: • Incorporating additional data sources: In addition to clinical signs, other data sources such as medical imaging, genetic data, and patient lifestyle information could be incorporated to enhance the classification accuracy. • Developing personalized models: Machine learning algorithms can be used to develop personalized classification models for individual patients, taking into account their unique characteristics and medical history. • Exploring other machine learning techniques: Different machine learning techniques, such as deep learning and ensemble methods, could be explored to further improve the performance of the classification models. JMEA Journal of Modern Educational Achievements Volume 5, 2024 https://scopusacademia.org/ 197 Overall, the proposed algorithm and software tool provide a valuable approach for improving the classification of diseases of the circulatory system. By leveraging machine learning techniques to identify informative signs and their relationships, our approach has the potential to enhance clinical decision-making and contribute to better patient outcomes. CONCLUSION The algorithm and software tool presented in this study provide a valuable approach for selecting informative signs and improving the classification of diseases of the circulatory system. They have the potential to enhance clinical decisionmaking and contribute to better patient outcomes. In conclusion, the algorithm and software tool presented in this study offer a novel and effective approach for classifying diseases of the circulatory system. By combining multiple informative signs and considering their relationships, our approach achieves higher accuracy and clinical interpretability compared to traditional methods. The automated software tool makes the complex selection process accessible to clinicians and researchers, enabling them to leverage machine learning techniques to improve patient care. Our findings have several implications for the management of circulatory system diseases: • Clinicians can use the identified complex of informative signs to more accurately diagnose and classify CVDs, leading to more targeted and effective treatment plans. • Researchers can use the software tool to explore different combinations of signs and develop personalized classification models for individual patients, taking into account their unique characteristics and medical history. • Healthcare organizations can adopt the software tool to standardize the classification process across different clinics and hospitals, ensuring consistency and reproducibility in the diagnosis and management of CVDs. Overall, our approach has the potential to significantly improve the accuracy and efficiency of CVD classification, ultimately leading to better patient outcomes and reduced healthcare costs. JMEA Journal of Modern Educational Achievements Volume 5, 2024 https://scopusacademia.org/ 198 REFERENCES 1. World Health Organization. Cardiovascular diseases (CVDs). https://www.who.int/news-room/fact-sheets/detail/cardiovascular-diseases-(cvds) 2. Vapnik, V. N. (1995). The nature of statistical learning theory. Springer Science & Business Media. 3. Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., ... & Va nderplas, J. (2011). Scikit-learn: Machine learning in Python. Journal of machine learning research, 12, 2825-2830.