How akramiot/wifi6_wifi7 works
File structure
• Top level: no separate folders—everything lives at root.
• Key files:
• WIFI7_mptcp.py – the main Python program containing data collection, ML preprocessing/training, and reporting logic.
• ML_PER_BER.docx/.pdf – design or research documentation describing the ML methodology and performance metrics (e.g., BER, throughput).
• Standard project metadata: .gitignore, LICENSE, README.md.
README
• Confirms: “ML based WIFI Performance analysis.”
Components and interactions
a. Main components
1. Data Collection Module
– Interfaces with OS/network stack (e.g., via subprocess calling tools like iw, iperf, or tcpdump).
2. Data Preprocessing & Feature Extraction
– Cleans raw logs, computes throughput, BER, RTT, jitter features.
3. ML Pipeline
– Model training, validation, and inference (classification/regression) using Python ML libraries.
4. Reporting & Visualization
– Generates plots, summary statistics, exportable reports (CSV, PDF, graphs).
b. Relationships & data flow
• Data Collection outputs raw measurement data → Preprocessing consumes raw data → Preprocessed features feed into the ML Pipeline → Model outputs performance predictions/analyses → Reporting consumes model outputs to produce visualizations and reports.
c. Architectural patterns
• Linear data‑pipeline pattern (ETL → ML → Reporting).
• Modular separation by function: each module can be a Python class or script function.
d. Technologies & frameworks
• Python 3.x
• Data libraries: pandas, NumPy
• ML frameworks: scikit‑learn (or TensorFlow/PyTorch if specified)
• Visualization: matplotlib, seaborn, or Plotly
• System interfacing: subprocess or OS modules
Elements
• Clear labels: “Data Collector,” “Preprocessor,” “ML Trainer/Predictor,” “Reporter”
• Blue rectangles for data at each stage
• Green for processing modules
• Orange for outputs/visuals