Data Science in Healthcare

Abstract: In this study, I analyzed the relationship between lifestyle and health factors regarding heart attack risks. Using a machine learning approach, I aimed to develop an effective prediction tool to support accurate early prevention of cardiovascular diseases.

1. Background

Heart attacks remain a leading cause of death globally. The primary challenge lies in the difficulty of detecting risks early based on complex lifestyle patterns. This research seeks to utilize public health data to evaluate the extent to which classification algorithms can accurately predict these risks.

2. Research Methodology

I utilized a dataset encompassing various medical and behavioral variables, then performed the following stages:

• Correlation Analysis: Identifying dominant factors such as cholesterol, obesity, and smoking habits.
• Model Experimentation: Implementing Gaussian Naïve Bayes (GNB) and Logistic Regression.
• Preprocessing: Data cleaning and splitting the dataset into training and testing sets (80:20).

3. Model Comparison

Based on the tests I conducted, the following results were obtained:

• Logistic Regression: Delivered the best performance with an accuracy of 70%. This model proved to be more stable in handling relationships between health variables.
• Gaussian Naïve Bayes: Provided an accuracy of 67%. Although slightly lower, this model is highly efficient in terms of processing time.

4. Key Findings

My analysis indicates that physical exercise has a significant negative correlation with risk (reducing the likelihood), while cholesterol levels and diabetes show a strong positive correlation with the probability of a heart attack.

5. Conclusion

This research confirms that the use of machine learning, specifically Logistic Regression, has great potential as an early screening tool in healthcare facilities. I recommend strict lifestyle management based on these predictive results to reduce mortality rates caused by heart attacks.