Pregledni rad
https://doi.org/10.32676/n.10.6
How to Profile Voters in Political Campaingns by Means of Artificial Intelligence?
Alen Kišić
; Zona Sjever d.o.o.
Sažetak
Artificial intelligence and machine learning are rapidly changing all spheres of society, including how
political campaigns work. This paper explores the application of two machine learning algorithms,
artificial neural networks and clustering, for voter profiling in political campaigns. The empirical
research is based on data about candidate activity in local elections in France, analyzing interactions
between candidates and voters through the social network Facebook. Data was analyzed for
candidates in 41 France cities France. The data included variables related to the number of shared
events, photographs, links, videos, statuses, and comments on content. Collecting data from social
networks is easier and cheaper than traditional methods of collecting public opinion. By applying
artificial neural networks, predictive models of election outcomes were created and identified
predictor variables of election results. The application of clustering algorithms using variables
that are the most significant predictors of elections enabled the segmentation of voters into more
homogeneous groups according to interests. The hybrid approach of predictive and descriptive
models based on machine learning algorithms enabled voter profiling. This study applies artificial intelligence and machine learning in politics for quality voter profiling based on social network data.
Ključne riječi
artificial intelligence, machine learning, predictive models, voters profiling
Hrčak ID:
324395
URI
Datum izdavanja:
19.12.2024.
Posjeta: 0 *