Postpartum Depression Classification

This document presents the classification results for a study on postpartum depression. Three different models were evaluated: Random Forest, Support Vector Machine (SVM), and Logistic Regression.

---

Model Evaluations

Random Forest

• Accuracy: 0.97
• Precision: Weighted average of 0.97
• Recall: Weighted average of 0.97
• F1 Score: Weighted average of 0.97

Confusion Matrix

Feature Importance

ROC Curve

Support Vector Machine (SVM)

• Accuracy: 0.74
• Precision: 0.73
• Recall: 0.74
• F1 Score: 0.73

Confusion Matrix

ROC Curve

Logistic Regression

• Accuracy: 0.74
• Precision: 0.73
• Recall: 0.74
• F1 Score: 0.73

Confusion Matrix

ROC Curve

---

Data Overview

The dataset comprises various attributes related to postpartum depression symptoms and behaviors. Here is a summary of the unique attributes:

• Age: ['35-40', '40-45', '30-35', '45-50', '25-30']
• Feeling sad or Tearful: ['Yes', 'No', 'Sometimes']
• Irritable towards baby & partner: ['Yes', 'No', 'Sometimes']
• Trouble sleeping at night: ['Two or more days a week', 'No', 'Yes']
• Problems concentrating or making decision: ['Yes', 'No', 'Often']
• Overeating or loss of appetite: ['Yes', 'No', 'Not at all']
• Feeling of guilt: ['No', 'Yes', 'Maybe']
• Problems of bonding with baby: ['Yes', 'Sometimes', 'No']
• Suicide attempt: ['Yes', 'No', 'Not interested to say']

Data provided by Md Parvez Mosaraf

---