Prototype Proposal for Profiling and Identification of TV Viewers using Watching Patterns

Aldis Erglis, Gundars Berzins, Irina Arhipova, Artis Alksnis, Evija Ansonska

2020

Abstract

Content based recommendation systems have widely been used for recommendations in ecommerce and in TV content recommendations for a long period of time. Such recommendation systems could help multimedia content providers separate content on individual level of TV viewers and offer better advertising options for media agencies and advertisers. One of the greatest challenges for providing individual TV content is identification of distinct TV viewers in household and link them with social economic and demographic metrics individually. From a technical point of view Machine Learning ensemble model should be created with several separate models for each need. In this study a prototype for a content-based recommendation system was created that can fulfil content targeting and watched content efficiency using real time watching data. The solution prototype covers all important parts of the model including data filtering, cleaning and transformation. The technical prototype allows to test efficiency of Machine Learning techniques used for prediction of household composition and social profiles assigned to an individual inhabitant of the household.

Paper Citation

in Harvard Style

Erglis A., Berzins G., Arhipova I., Alksnis A. and Ansonska E. (2020). Prototype Proposal for Profiling and Identification of TV Viewers using Watching Patterns.In Proceedings of the 22nd International Conference on Enterprise Information Systems - Volume 1: ICEIS, ISBN 978-989-758-423-7, pages 571-578. DOI: 10.5220/0009458805710578

in Bibtex Style

@conference{iceis20,
author={Aldis Erglis and Gundars Berzins and Irina Arhipova and Artis Alksnis and Evija Ansonska},
title={Prototype Proposal for Profiling and Identification of TV Viewers using Watching Patterns},
booktitle={Proceedings of the 22nd International Conference on Enterprise Information Systems - Volume 1: ICEIS,},
year={2020},
pages={571-578},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0009458805710578},
isbn={978-989-758-423-7},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 22nd International Conference on Enterprise Information Systems - Volume 1: ICEIS,
TI - Prototype Proposal for Profiling and Identification of TV Viewers using Watching Patterns
SN - 978-989-758-423-7
AU - Erglis A.
AU - Berzins G.
AU - Arhipova I.
AU - Alksnis A.
AU - Ansonska E.
PY - 2020
SP - 571
EP - 578
DO - 10.5220/0009458805710578