Machine Learning & Artificial Intelligence
Today’s telecom networks create massive amounts of valuable data. In fact, ‘Big Data’ has become so big that it requires the use of sophisticated tools and algorithms to extract its full potential. Artificial intelligence (AI) and machine-learning (ML) have emerged as the data analytics fields of expertise that can cope with this complexity while providing carriers with incredible new insights and avenues for growth.
5G networks and new IoT services will depend on ML and AI to enable the scalability, automation and network intelligence required for success. At TEOCO, our team of data scientists have developed advanced machine learning capabilities that leverage our long history and subject matter expertise in the telecom industry.
The following are a few examples of how we are leveraging ML/AI to solve challenges for network operators:
Machine-Learning Root Cause Analysis (ML-RCA)
Helix Analytics uses the latest machine learning techniques for automating the screening and root cause analysis of service assurance alarms. Locating the source of problems using our patented algorithms that analyze historical and current alarms, and the behavior of network and service resources. ML-RCA automatically suggests clustering and correlation between alarms and points to the potential root-cause.
Helix Analytics and Helix Performance Management utilize machine learning for creating alarm thresholds that are optimized for their environment. The Adaptive Thresholds tool automates how thresholds are created and maintained, significantly reducing the number required. The tool uses several complex statistical algorithms for calibrating and fine-tuning the early detection of non-trivial problems, resulting in faster time to market for new services.
As part of Helix Analytics and Helix Performance Management solutions, the Anomaly Detection tool discovers irregular patterns by automatically identifying high levels of abnormal KPI behavior, or non-trivial KPI-to-KPI relationships.
Our VoLTE Analytics solution, INsync, leverages the latest in machine learning to optimize the call experience. For example, the tool has an algorithm that assigns an individual MOS (measure of success) score to each of the billions of VoLTE calls being analyzed. This enables network engineers and care staff to easily identify and troubleshoot issues that result in low scores.
Encrypted Video – delivery of quality video content
With the explosion of streaming video, the need to understand the quality of the user experience is paramount. Through investments into machine learning, our INsync Video Analytics solution measures video performance by detecting encrypted video characteristics in near real time.
App Detection – increasing subscribers & revenue
TEOCO’s Automatic OTT App Detection tool identifies encrypted and non-encrypted apps on a per-user basis. By identifying which categories of Apps are being used by subscribers, along with when and where, CSPs can better understand subscriber behavior and how their networks are being utilized – in order to deliver a better customer experience. These insights can also be monetized by CSPs.
Churn Prediction – retaining subscribers
TEOCO’s Prescriptive Analytics tool utilizes subscriber data – such as rate plans and account activity, paired with network data – to help accurately predict and reduce subscriber churn.
Financial Cost Audits – cost reduction
Our team of financial auditors and Revenue Assurance subject matter experts train machine learning software that quickly identifies many cost savings opportunities.
RPA – robotic process automation
TEOCO utilizes BOTs within our Cost Management and Revenue Assurance solutions to improve business efficiencies and eliminate human errors. BOTs free up valuable resources to work on higher-value tasks. Powered by Machine Learning, NLP (natural language processing) and rule-based algorithms, TEOCO’s RPA tools reduce the need for time consuming tasks with the latest cutting-edge technology.
Geolocation – indoor/outdoor call classification
The Mentor geolocation algorithm assigns an indoor or outdoor classification to calls based on data analysis and unsupervised machine learning principles. This allows engineers to filter data and gain a clear picture of indoor quality of experience as well as outdoor quality of experience. With this data they are able to make more informed decisions on what optimizations would most improve the network.
Geolocation – Call height classification
The Mentor geolocation algorithm is enriched with building height data, enabling unique techniques based on supervised and unsupervised machine learning principals for geolocating calls within high-rise buildings. An engineer can then select floor classifications in Mentor’s location profile filters for more focused analysis and troubleshooting of traffic on lower, middle or upper floors of buildings.
Automatic Problem Detection Clustering
Mentor’s Automatic Problem Detection uses machine learning clustering algorithms to create geographical polygons highlighting areas where KPIs have been breached. Detected problems are assigned a severity level, enabling engineers to focus on the most severe service-affecting issues first.
How can we help?
Whether you are looking to improve your planning and optimization capabilities or gain greater analytical insights into your subscribers, TEOCO’s ML and AI powered products can help. Contact us to learn more.