| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 2.82 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
After an experience with a service or product, consumers tend to share their impressions
online. For companies is crucial to keep track of this information by collecting and analyzing
it. Due to the large volume, unstructured nature, and emotional polarity, online reviews
require automatic analytical treatment. Text mining, topic modelling, and sentiment analysis
are challenging approaches to handle the need to obtain knowledge from online reviews.
However, in this context, questions may arise regarding the efficiency of the automatic
interpretation of online reviews. The main objective of this study is to understand if the effect
of TAK (The Anna Karenina) principle is observed in online reviews of Portuguese restaurants
collected from TripAdvisor. We used methods such as Latent Dirichlet Allocation and
Sentiment Analysis. TAK principle is applied in different areas, so on the one hand it can be
considered universal for any context. Nevertheless, it is important to note that the principle
was formed recently, and it is only at the beginning of its path of application. Further studies
are needed to test whether TAK principle is verified in various contexts and if it is confirmed
it would mean that the TAK limits the application of topical analysis when the nature of the
data has great variability, that is, mostly customer dissatisfaction topics. This study contributes
to important discoveries that have managerial implications in the sense of proposing specific
strategies and approaches that can improve the customer experience, and theoretically by
contributing to the universality of TAK principle.
Descrição
Dissertation presented as the partial requirement for obtaining a Master's degree in Information Management, specialization in Knowledge Management and Business Intelligence
Palavras-chave
Anna Karenina Principle Topic Modelling Sentiment Analysis Portuguese restaurants TripAdvisor
