Kouotou/telcoqoepredict

Using different Machine Learning techiques to predict QoE from QoS KPIs in the Telecommunication space of Cameroon # QoE Prediction from QoS Parameters in Cameroon Telecommunication Networks **Master of Science in Telecommunications and Network Engineering Project** *University of Buea, Cameroon* --- ## 📌 Project Overview This research-grade machine learning project aims to predict **Quality of Experience (QoE)** based on **Quality of Service (QoS)** key performance indicators (KPIs) collected from telecommunication networks in Cameroon. By moving beyond simple network metrics (like latency and packet loss) and predicting actual perceived user experience, this project helps network engineers prioritize traffic, diagnose bottlenecks, and optimize resource allocation effectively. The repository includes: 1. **A highly documented Jupyter Notebook** covering Exploratory Data Analysis (EDA), feature engineering, model training, and Explainable AI (SHAP). 2. **An Interactive Streamlit Dashboard** allowing real-time scenario simulation and batch predictions for network administrators. --- ## 🚀 Features - **Data Preprocessing Pipeline:** Robust handling of malformed telecom data, missing values, and outliers. - **Feature Engineering:** Advanced telecom KPIs (e.g., `Throughput_Ratio`, `Network_Instability_Index`). - **Machine Learning Models:** Comparisons across Linear models, Random Forest, XGBoost, CatBoost, LightGBM, and SVR. - **Explainable AI (SHAP):** Transparent predictions showing exactly how Jitter, Latency, and Packet Loss affect a user's MOS (Mean Opinion Score). - **Streamlit Dashboard:** A production-ready UI for real-time QoE simulations. --- ## ⚙️ Installation Guide ### Prerequisites Make sure you have **Python 3.10+** installed on your machine. ### 1. Clone the repository ```bash git clone github.com cd QoePredictModel ``` ### 2. Create a Virtual Environment It is highly recommended to use a virtual environment to manage dependencies. **On Windows:** ```powershell python -m venv qoepredict .\qoepredict\Scripts\activ …