In this talk, We present a methodology to predict network degradation in the telecom sector. We will be explaining how to forecast degradation of network key performance indicators (KPIs) and providing (24 Hrs. in advance) alerts to network operations team to take preemptive actions before degradation affects network performance

Outline

 Data collected from various sensors such as network counters, traffic (Voice & Data), alarm occurrences in the network etc. were loaded into Spark from SQL server directly. Data was aggregated at every hour.
 Data was treated to impute missing values, outliers etc
 Feature engineering was done and additional variables were brought in
 Used sophisticated and advanced predictive models such as Random Forest, GLM, SVM, Neural network & Decision Tree and ensembled the results from the above models to get more accurate prediction

Speaker bio

Roshni is an analytics & machine learning enthusiast with experience of close to 10 years having worked with clients across domains including technology , financial services, retail and manufacturing. In previous assignment, she also handled training sessions in Analytics and SAS for individuals and some of the corporates in India.

A Mathematics major and MBA from Symbiosis Pune in Marketing she also has completed a formal training from Indian Institute of Management in Business Analytics and Intelligence