Mini-project
Electronics, Semiconductors & Telecom
Network fault severity predictor
Predict how severe a network disruption will be from location and event logs.
The brief
A telecom operator wants to know which reported disruptions will turn into major outages. Predict fault severity at each location from event types, log features and resource types, and show which signals matter most. Good looks like: sensible feature engineering from logs, a multi-class model with log-loss, and clear drivers of severe faults.
What you will build
- Join the location, event, log and resource tables
- Engineer features per location
- Train and compare multi-class models
- Evaluate with log-loss and per-class recall
- Explain the drivers of severe faults
Free, open data to build it with
Check each dataset's licence before you publish your work. Your mentor gives your team its own version of the brief (data, scope or region), so every team's project is its own.