# Analyzing the impact of feature selection on the accuracy of heart disease prediction

**Authors:** Muhammad Salman Pathan, Avishek Nag, Muhammad Mohisn Pathan, Soumyabrata Dev  
**arXiv:** [2206.03239](https://arxiv.org/abs/2206.03239)  
**Date:** 2022-06-07  
**Summary page:** https://www.summarizepaper.com/en/paper/2206.03239v1/

## Key points

- Heart disease is a major global health concern with significant mortality rates
- Accurate and timely diagnosis is crucial for preventing further damage to patients
- Non-invasive medical procedures based on AI and ML have shown promise in diagnosing heart disease accurately and quickly
- Predicting heart disease is complex due to large datasets and intricate feature relations
- A research study aimed to identify important risk factors from datasets to aid in accurate classification of heart disease with fewer complications
- Relevant features had a significant impact on the accuracy of heart disease classification
- Using a reduced number of features improved the performance of classification models compared to using the full feature set
- Training time was reduced without compromising accuracy
- Feature selection can improve heart disease prediction accuracy, reduce complexity, and training time
- Identifying key risk factors enhances healthcare professionals' ability to diagnose and treat heart disease effectively.

## Summary

Heart disease has become a major global health concern in the past decade, leading to significant mortality rates. Accurate and timely diagnosis is crucial for preventing further damage to patients. In recent years, non-invasive medical procedures based on artificial intelligence (AI) and machine learning (ML) have shown promise in accurately diagnosing heart disease in less time. However, predicting heart disease is a complex task due to the increasing size of medical datasets and the need to understand intricate feature relations. To address this challenge, a research study conducted by Muhammad Salman Pathan et al aimed to identify the most important risk factors from highly dimensional datasets that could aid in the accurate classification of heart disease with fewer complications. The researchers analyzed two heart disease datasets with various medical features to gain a broader understanding. The results of their analysis demonstrated that relevant features had a significant impact on the accuracy of heart disease classification. Even when using a reduced number of features, the performance of classification models improved significantly compared to models trained on the full feature set. Additionally, training time was reduced without compromising accuracy. This research highlights the potential benefits of feature selection in improving heart disease prediction accuracy while reducing complexity and training time. By identifying key risk factors, healthcare professionals can enhance their ability to diagnose and treat heart disease effectively.