Master Thesis - Network KPI Time Series Prediction using Bayesian Non-Parametric Method

Job description

Job Description

Date: Oct 22, 2018

Background

The Radio System Management Analytics section within the Product Development Unit 4G & 5G is responsible for driving the Radio Network analytics operation at PDU 4G & 5G RAN. The scope ranges from setting the strategic direction and developing Machine Learning and AI solutions in the RAN product to using analytics techniques for trouble shooting and understanding the product behavior in field. We are now offering a Thesis position targeting first half of 2019.

Key Performance Indicators (KPIs) are a set of critical metrics of the performance cellular networks. KPI prediction is a key task for 4G and 5G networks with critical applications such as identifying security threats, monitoring, troubleshooting, and predicting service interrupts due to irregular usage patterns or gradual failures. Real-time KPI prediction may help improving the network performance as well as costumer experience. Classical prediction methods such as Auto-Regressive Moving Average (ARMA) include finite model parameters which may work well for short-term and stationary time-series prediction. Applicability of these methods is limited in real-network KPI prediction in which time-series are often huge in size and show certain non-stationary behavior.

A Bayesian non-parametric method, on the other hand, provides an effective tool to long-term time-series prediction, where the size of models are allowed to grow with the data size. Rather than comparing models that vary in complexity, the Bayesian non-parametric approach is to fit a single model that can adapt its complexity to the data. Furthermore, the actual forecast from the Gaussian processes is supported by confidence intervals, making this model particularly interesting for KPI prediction.

Thesis description

The aim of this thesis is to develop a general KPI prediction framework based on Bayesian non-parametric approach for Key Performance Indicator (KPI) prediction. This involves studying and understanding the Bayesian non-parametric approach as well as implementing the method on data collected from live networks. Comparison with the classical prediction methods (e.g., ARMA model) is also needs to be investigated. The prediction framework then will be evaluated in one or more radio network use cases.

Qualifications

You should be self-motivated and used to working with others in project teams. The positions also require you to be fluent in English, both written and spoken. In return, you will get to perform your thesis work with cutting-edge technology in a stimulating learning environment with a friendly atmosphere.

Key qualifications:

·  Background in Engineering, Computer Science, Telecommunications or similar
·  Significant part of courses associated with the degree are accomplished
·  Strong analytical skills and ability to acquire new knowledge and apply it in the job
·  Ability to formulate problems and solve them independently and with the team
·  Basic knowledge and experience in the probabilistic modeling, Gaussian processes, time series analysis
·  Very good programming skills (Java/Python/R)
·  Basic understanding of mobile networks would be a merit
·  Willingness to write publication

We will provide active support for the motivated thesis worker in the work to successfully fulfil the requirements for the Master Thesis.

Additional Details

The work is proposed to start in January 2019. The work is for one student to be performed during first half of 2019. Location is at Ericsson in Stockholm (Kista), Sweden.

For informal queries, feel free to email Sholeh Yasini at or Ulf Norholm at

Please submit your application in English as soon as possible - we are working continuously with candidate selection.

Ericsson provides equal employment opportunities (EEO) to all employees and applicants for employment without regard to race, color, religion, sex, sexual orientation, marital status, pregnancy, parental status, national origin, ethnic background, age, disability, political opinion, social status, veteran status, union membership or genetics.

Ericsson complies with applicable country, state and all local laws governing nondiscrimination in employment in every location across the world in which the company has facilities. In addition, Ericsson supports the UN Guiding Principles for Business and Human Rights and the United Nations Global Compact.

This policy applies to all terms and conditions of employment, including recruiting, hiring, placement, promotion, termination, layoff, recall, transfer, leaves of absence, compensation, training and development.

Ericsson expressly prohibits any form of workplace harassment based on race, color, religion, sex, sexual orientation, marital status, pregnancy, parental status, national origin, ethnic background, age, disability, political opinion, social status, veteran status, union membership or genetic information.

Primary country and city: Sweden (SE) || || Stockholm || Stud&YP

Req ID: 261478