In time, electronic Word Of Mouth has become a resource to support the decision-making process. Different techniques have been proposed to extract information from online textual data. We propose a semi-supervised clustering model able to identify clusters homogeneous with respect to the overall sentiment of the analyzed texts. The model is built by combing Sentiment Analysis, and Network-based Semi-supervised Clustering. We apply the model to the Booking.com data related to the Sardinian hotels. The first results highlight the presence of different clusters non-overlapped in terms of the distribution of the overall sentiment.
Nel tempo, il passaparola online e diventato una risorsa a supporto del`processo decisionale. Sono state proposte diverse tecniche per estrarre informazioni da dati testuali online. Proponiamo un modello di clustering semi supervisionato in grado di identificare cluster omogenei rispetto al sentiment complessivo dei testi analizzati. Il modello e costruito combinando diverse metodologie come Sentiment Analysis e Network-based Semi supervised Clustering. Applichiamo il modello ai dati di Booking.com relativi agli hotel che operano in Sardegna. I primi risultati evidenziano la presenza di diversi cluster non sovrapposti con riferimento alla distribuzione del sentiment complessivo.
A semi-supervised clustering method to extract information from the electronic Word Of Mouth
Contu, Giulia;Frigau, Luca;Romano, Maurizio;Ortu, Marco
2022-01-01
Abstract
In time, electronic Word Of Mouth has become a resource to support the decision-making process. Different techniques have been proposed to extract information from online textual data. We propose a semi-supervised clustering model able to identify clusters homogeneous with respect to the overall sentiment of the analyzed texts. The model is built by combing Sentiment Analysis, and Network-based Semi-supervised Clustering. We apply the model to the Booking.com data related to the Sardinian hotels. The first results highlight the presence of different clusters non-overlapped in terms of the distribution of the overall sentiment.| File | Dimensione | Formato | |
|---|---|---|---|
|
A semi-supervised clustering method to extract information from the electronic Word Of Mouth.pdf
Solo gestori archivio
Tipologia:
versione editoriale (VoR)
Dimensione
608.1 kB
Formato
Adobe PDF
|
608.1 kB | Adobe PDF | Visualizza/Apri Richiedi una copia |
I metadati presenti in IRIS UNICA sono rilasciati con licenza Creative Commons CC0 1.0 Universal, mentre i file delle pubblicazioni sono protetti da diritto d'autore, salvo diversa indicazione.



