Napredna pretraga

Pregled bibliografske jedinice broj: 819179

Revealing the structure of domain specific tweets via complex networks analysis


Močibob, Edvin; Martinčić-Ipšić, Sanda; Meštrović Ana
Revealing the structure of domain specific tweets via complex networks analysis // MIPRO, SP, 2016 / Biljanović, Petar (ur.).
Opatija, Croatia, 2016. str. 1904-1908 (predavanje, međunarodna recenzija, cjeloviti rad (in extenso), znanstveni)


Naslov
Revealing the structure of domain specific tweets via complex networks analysis

Autori
Močibob, Edvin ; Martinčić-Ipšić, Sanda ; Meštrović Ana

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
MIPRO, SP, 2016 / Biljanović, Petar - Opatija, Croatia, 2016, 1904-1908

ISBN
978-953-233-087-8

Skup
MIPRO 39th International Convention, SP. IEEE

Mjesto i datum
Opatija, Croatia, 30.5-03.6.2016

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
Weighted complex networks; tweets; link prediction; tweet polarity
(Eighted complex networks; tweets; link prediction; tweet polarity)

Sažetak
In this paper we explore the relation between different groups of tweets using complex network analysis and link prediction. The tweets were collected via the Twitter API depending on their textual content. That is, we searched for the tweets in English language containing specific predefined keywords from different domains. From the gathered tweets a complex network of words was formed as a weighted network. Nodes represent words and a link between two nodes exists if these two words co-occur in the same tweet, while weight denotes the co-occurrence frequency. The Twitter search was repeated for four different search criteria (API queries based on different tweet keywords), thus resulting in four networks with different nodes and links. The resulting networks were subjects to further network analysis, as comparison of numerical properties for different networks and link prediction for individual networks. This paper shows the tweet scraping process, our approach to building the networks, the measures we calculated for them, the differences and similarities between different networks we built and our success in predicting future links.

Izvorni jezik
Engleski

Znanstvena područja
Računarstvo, Informacijske i komunikacijske znanosti



POVEZANOST RADA


Ustanove
Sveučilište u Rijeci - Odjel za informatiku