Catalyst Project
Glaucus-Precision on Telco Data
Winner! Digital Transformation Asia 2019: Outstanding Catalyst – business impact
In traditional wireless network operation and maintenance, silo data evaluation standards and interfaces, complicated cell situations, and manual troubleshooting has resulted in complex root cause analysis and delays in recovery. To deal with this challenge, our project explores the following solutions:
- Using unique data quality control technology, standardized data cleansing and quality analysis to process data from multi-vendor networks, making sure it is ready for AI model training;
- Based on the cell situation and distribution of KPIs, an AI model that combines single and multi-dimension indicators is selected, after evaluating multiple algorithms and in combination with experts’ input. An accurate anomaly detection algorithm without fixed threshold is executed on cells to discover the possible network capacity problem and congestion situations;
- For the detected cell congestion problem, long-term and short-term AI forecast model is established using the automated hyper-parameter selection technology of AutoML. Thus, cell capacity can be predicted and a proposal for capacity expansion will be given.
Shortlisted
- Best new Catalyst in show
- Outstanding contribution to TM Forum assets
Champions
Participants
Videos & Images
Glaucus-Precision on Telco Data – Interview
Winners of Catalyst Team Award – Outstanding Catalyst – business impact