IncelliAna


Insights & Predictions

This module of Incelligent software framework builds analyzes data sources and extracts valuable knowledge that exists in hidden correlations. It builds insights, indicatively regarding the quality and efficiency of the network service, and foresights (predictions), regarding the situations to be addressed in time and space in the near/mid-term future.


BigData medium One of the main innovations, in order to build the insights/foresights, is the exploitation of highly heterogeneous input. In general, input is either telecommunication-related or non-telecommunication-related. In our various studies and pilots the telecommunication related input includes aspects like the traffic load offered, the network capacity / coverage / performance metrics, the QoS delivered, etc. The collection relies on standardized interfaces of network management systems. The non-telecommunication-related input includes aspects like the weather, the date, the associated events in time and space, the user mobility, marketing actions, etc.

Our product uses advanced machine learning and predictive analytic mechanisms for exploiting the heterogeneous input. Our output offers simple, intuitive and sophisticated visualizations. Essentially, these express the insights/foresights in the following sample forms:

"probability x% that a cell will encounter y% load level at specific time / date range"

"probability x% that cells {a,b} will be congested on Tuesday evening"

"probability x% that a cell will be underutilized between hh:mm and hh:mm"