Early stage Hypertension Stage Prediction

Overview

The Hypertension Stage Prediction Project utilizes a decision tree classifier (DTC) trained on a comprehensive hypertension dataset. This dataset includes critical features such as blood pressure readings, BMI, cholesterol levels, and lifestyle indicators. The primary objective is to predict hypertension stages, enabling early identification and intervention for individuals at risk.

The dataset includes preprocessed features such as:

• Age (years): Converted from raw 5-digit age format.
• BMI: Calculated using height and weight.
• Blood Pressure Categories: Encoded for efficient processing.

Model: The project uses a DecisionTreeClassifier, achieving 100% accuracy on the test dataset after normalization using MinMaxScaler. This ensures feature scaling for optimal model performance.

Agile Features (Implemented in Sprints)

Sprint 1: Data Collection and Preprocessing

• Dataset cleaning: Removed irrelevant columns (id, bp_category).
• Applied Ordinal Encoding to categorical variables (e.g., bp_category_encoded).
• Feature engineering: Added age_years and bmi.

Sprint 2: Model Training and Evaluation

• Split dataset: 80-20 training-test split with a fixed random state (101).
• Applied MinMaxScaler for feature scaling.
• Achieved 100% test accuracy using a DecisionTreeClassifier.

Sprint 3: Integration and Deployment

• Developed an interactive HTML form for feature input.
• Backend implementation in Flask to handle predictions and return results in real-time.

Results

The model categorizes hypertension into the following stages:

1. Elevated
2. Hypertension Stage 1
3. Hypertension Stage 2
4. Normal

Results are delivered with near-instantaneous processing, providing actionable insights to users. Testing demonstrated flawless stage prediction on unseen data.

Conclusion

Use Cases:

1. Clinical Decision Support: Assists healthcare providers in classifying hypertension stages for diagnosis and treatment planning.
2. Health Monitoring Apps: Integration into wearable devices or mobile applications for proactive health management.
3. Public Health Analytics: Enables trend analysis in population health studies.