Managing and Optimising IoT Data and ML applications dependencies

Loading...
Thumbnail Image

URL

Journal Title

Journal ISSN

Volume Title

Perustieteiden korkeakoulu | Master's thesis

Date

2020-05-18

Department

Major/Subject

Computer Science

Mcode

SCI3042

Degree programme

Master’s Programme in Computer, Communication and Information Sciences

Language

en

Pages

64

Series

Abstract

The thesis is motivated by applying Machine Learning and analytical applications over IoT data in the production environment. We focus on how the two fields work together in production. Specifically, the research is done towards dependencies between IoT data and ML applications, and how to manage and optimise these dependencies. The aim of the thesis was to create a framework for managing dependencies as well as produce an implementation of the framework so that Helvar could utilise it for their benefit. First, we look into the background of the problem and environment as where and how the thesis was done. Next, state-of-the-art solutions are reviewed for managing IoT data and ML applications in production. Afterwards, dependencies of the two fields are narrowed down to select few. The framework was designed for managing dependencies by utilising and slightly adjusting state-of-the-art solutions. The evaluation of the framework was done by implementing a proof of concept pipeline that takes IoT data and utilises ML to produce predictions on air quality in the building. The implementation and design of the framework proved to be useful for the company at which the thesis was done. Therefore, it shows that framework can help data engineer with optimisation and management of IoT data and ML applications dependencies.

Description

Supervisor

Truong, Hong-Linh

Thesis advisor

Nasir, Omar

Keywords

internet of things, machine learning, cloud, Amazon

Other note

Citation