jomarie04/mobile-app-churn-predictor

---

model_name: mobile-app-churn-predictor license: apache-2.0 language: en tags: - tabular - churn-prediction - user-analytics - classification - synthetic-data metrics: - accuracy - roc_auc

Mobile App Churn Predictor

Model Overview

Mobile App Churn Predictor is a machine learning model designed to predict whether a user is likely to churn based on app usage behavior.

The model is trained on synthetic mobile application analytics data and is suitable for demos, tutorials, and experimentation.

---

Model Details

• Model Name: mobile-app-churn-predictor
• Model Type: Binary Classification
• Framework: scikit-learn
• Input: Tabular CSV data
• Output: Churn probability (Yes / No)

---

Training Data

Synthetic mobile app usage dataset with the following features:

Feature	Type	Description
age	Integer	User age
device_os	Categorical	Android / iOS
daily_active_minutes	Integer	Avg daily usage
sessions_per_day	Integer	App sessions
features_used	Integer	Number of features used
days_since_install	Integer	Days since app install
churn	Binary	Target label

---

Intended Use

✅ Churn prediction demos
✅ User behavior modeling
✅ Hugging Face Spaces examples
❌ Production decision-making

---

Evaluation Results

Metric	Score
Accuracy	0.88
ROC-AUC	0.91

---

Limitations

• Synthetic data only
• Simplified user behavior
• Not representative of real-world scale

---

Ethical Considerations

This model does not use real personal data.
Care should be taken when applying churn models to avoid unfair targeting.

---

License

This model is released under the Apache License 2.0.