Search results

The purpose of this study is to detect these reviews’ complex emotions, visualize and analyze them. Movie reviewers’ moviescores and reviews can be analyzed with respect to their emotion content, aggregated and projected onto a movie, resulting in an emotion map for a movie. It is then possible for a moviegoer to choose a movie, not only on the basis of movie scores and reviews, but also on the basis of aggregated emotional outcome of a movie as reflected by its emotion map displaying certain emotion map patterns desirable for the moviegoer.

The authors use the hourglass of emotion model to find the emotional scores of words of a review, then they use singular value decomposition to reduce the data dimension into singular scores. Once, they have the emotional scores of reviews, the authors cluster them by using k-means algorithm to find similar emotional levels of movies. Finally, the authors use heat maps to visualize four dimensions in a figure.

The authors are able to find the emotional levels of movie reviews, represent them in single scores and visualize them. The authors look the similarities and dissimilarities of movies based on their genre, ranking and emotional statuses. They also find the closest emotion levels of movies to a given movie.

The authors detect complex emotions from the text and simply visualize them.

The purpose of this paper is to combine basic movie information factors, external factors and review factors, to predict box-office performance and identify the most crucial factor of influence for box-office performance.

Five movie genres and first-week movie reviews found on IMDb were collected. The movie reviews were quantified using sentiment analysis tools SentiStrength and Stanford CoreNLP, in which quantified data were combined with basic movie information and external environment factors to predict movie box-office performance. A movie box-office performance prediction model was then developed using data mining (DM) technologies with M5 model trees (M5P), linear regression (LR) and support vector regression (SVR), after which movie box-office performance predictions were made.

The results of this paper showed that the inclusion of movie reviews generated more accurate prediction results. Concerning movie review-related factors, the one that exhibited the greatest effect on box-office performance was the number of movie reviews made, whereas movie review content only displayed an effect on box-office performance for specific movie genres.

Because this paper collected movie data from the IMDb, the data were limited and primarily consisted of movies released in the USA; data pertaining to less popular movies or those released outside of the USA were, thus, insufficient.

This paper helps to verify whether the consideration of the features extracted from movie reviews can improve the performance of movie box-office.

Through various DM technologies, this paper shows that movie reviews enhanced the accuracy of box-office performance predictions and the content of movie reviews has an effect on box-office performance.

Electronic word-of-mouth (eWOM) is very important for consumer decision making; previous international product diffusion studies have investigated eWOM and cultural factors that influence consumers’ acceptance of new products, but they have not adequately compared the differences in these factors between the USA and China. Therefore, the purpose of this paper is to compare the impact of eWOM on consumer choices in China and the USA. The authors addressed the following questions: What are the cross-cultural differences in consumers’ eWOM behavior between the USA and China: Which genres of Hollywood movies have better cross-culture predictability in terms of box office performance; and What factors affect the success of Hollywood movies in entering the Chinese market?

Real eWOM data were collected from two online movie review websites, IMDb.com (the USA) and Douban.com (China), from January 2010 to December 2015. In addition, box office revenue information was collected from BoxOfficeMojo.com. The authors used an independent sample t-test to check whether the differences in consumers’ eWOM behavior between China and the USA and different types of movie lead to cultural discount differences. Furthermore, a log-linear regression model is used to examine which factors influence the commercial success of new movies.

There are specific similarities and differences between the American and Chinese movie markets. First, the results show that American consumers are more engaged in online review systems and tend to submit extreme reviews, but Chinese consumers tend to submit moderate reviews on movies, and the eWOM variance there is smaller than in the USA. Second, genres are useful variables as indicators of movie content; the genres of comedy and drama are not popular in the Chinese market. Finally, eWOM variance has a positive impact on box office in China, but eWOM variance has no impact on the US box office. In addition, the interactive effect of the average rating and eWOM variance on sales is positively significant in China. Importantly, the one-star reviews have a negative impact on the Chinese box office, but it has no impact on US box office.

Understanding how cultural factors influence consumer eWOM communication will help managers to better apply this new marketing communication tool to create more aggressive and targeted promotional plans. Marketers may use eWOM behavior to better respond to and target consumers to overcome barriers to the selection of their products by consumers. Therefore, more effective management of eWOM can improve the acceptance of and preference for products in different cultural consumer groups.

This study expands the existing body of knowledge on eWOM and international marketing literature. Clearly, culture is an important determinant of eWOM’s impact on sales. In addition, it provides strategic direction and practical implications for eWOM communication management in cross-cultural settings.

The lighthouse tourism, which has been flourishing in several coastal areas and port cities with waterfront, provides the ideal scenario for escape experiences. The purpose of this paper is to discuss the implicit (dark tourism) meanings, symbolisms and emotions evoked by lighthouses, in particular those related with recreational storm chasing, “land’s ends” pilgrimage and gaze upon dystopic places.

Using a qualitative approach such as filmography’s content analysis (filtered by IMDb database), photo elicitation and engagement with lighthouses promotion websites, this study searched for evidences supporting the classification of lighthouse tourism as a “lighter” dark tourism product.

The qualitative information gathered from different sources provided support for a taxonomy of motives for engaging (dark) lighthouse experiences: risk recreation; isolation and loneliness; pilgrimage; shipwrecking; memorials; dystopia; and gaze for “ice palaces.”

This conceptual paper suggests a taxonomy for a systematic classification of dark lighthouse experiences and suggested some research propositions for further research.

Public decision makers, maritime authorities and tourism operators may acknowledge the theoretical and practical contributions provided by this paper and develop innovative escape experiences.

The lighthouse tourism is an innovative and creative way to promote the sustainable development of waterfronts of port cities, giving more “energy” to these coastal and often peripheral areas.

The paper fills a gap in the literature that so far never had deeply explored the relationship between the lighthouses’ meanings/experiences and dark tourism and introduces the innovative concept of (dark) lighthouse tourism.

The purpose of this paper is to expose librarians, scholars and other interested parties to the numerous films available concerning the 1989 and 1991 European revolutions. The films that are discussed can potentially be used as ancillary sources that will lead to a more in-depth understanding of these topics.

This paper is a literature review examining films relating to the 1989 and 1991 revolutions in Eastern Europe and the former Soviet Union. The findings are presented in the form of an annotated bibliography.

A total of 24 films from eight countries are presented in this annotated bibliography.

In researching this paper, the authors have been unable to find any similar works, which makes this work of particular value to those wanting to learn more about this period of change in Eastern Europe and the former Soviet Union.

This paper aims to propose a statistical and context-aware feature reduction algorithm that improves sentiment classification accuracy. Classification of reviews with different granularities in two classes of reviews with negative and positive polarities is among the objectives of sentiment analysis. One of the major issues in sentiment analysis is feature engineering while it severely affects time complexity and accuracy of sentiment classification.

In this paper, a feature reduction method is proposed that uses context-based knowledge as well as synset statistical knowledge. To do so, one-dimensional presentation proposed for SentiWordNet calculates statistical knowledge that involves polarity concentration and variation tendency for each synset. Feature reduction involves two phases. In the first phase, features that combine semantic and statistical similarity conditions are put in the same cluster. In the second phase, features are ranked and then the features which are given lower ranks are eliminated. The experiments are conducted by support vector machine (SVM), naive Bayes (NB), decision tree (DT) and k-nearest neighbors (KNN) algorithms to classify the vectors of the unigram and bigram features in two classes of positive or negative sentiments.

The results showed that the applied clustering algorithm reduces SentiWordNet synset to less than half which reduced the size of the feature vector by less than half. In addition, the accuracy of sentiment classification is improved by at least 1.5 per cent.

The presented feature reduction method is the first use of the synset clustering for feature reduction. In this paper features reduction algorithm, first aggregates the similar features into clusters then eliminates unsatisfactory cluster.

A community demonstrates the unique qualities and relationships between its members that distinguish it from other communities within a network. Network analysis relies heavily on community detection. Despite the traditional spectral clustering and statistical inference methods, deep learning techniques for community detection have grown in popularity due to their ease of processing high-dimensional network data. Graph convolutional neural networks (GCNNs) have received much attention recently and have developed into a potential and ubiquitous method for directly detecting communities on graphs. Inspired by the promising results of graph convolutional networks (GCNs) in analyzing graph structure data, a novel community graph convolutional network (CommunityGCN) as a semi-supervised node classification model has been proposed and compared with recent baseline methods graph attention network (GAT), GCN-based technique for unsupervised community detection and Markov random fields combined with graph convolutional network (MRFasGCN).

This work presents the method for identifying communities that combines the notion of node classification via message passing with the architecture of a semi-supervised graph neural network. Six benchmark datasets, namely, Cora, CiteSeer, ACM, Karate, IMDB and Facebook, have been used in the experimentation.

In the first set of experiments, the scaled normalized average matrix of all neighbor's features including the node itself was obtained, followed by obtaining the weighted average matrix of low-dimensional nodes. In the second set of experiments, the average weighted matrix was forwarded to the GCN with two layers and the activation function for predicting the node class was applied. The results demonstrate that node classification with GCN can improve the performance of identifying communities on graph datasets.

The experiment reveals that the CommunityGCN approach has given better results with accuracy, normalized mutual information, F1 and modularity scores of 91.26, 79.9, 92.58 and 70.5 per cent, respectively, for detecting communities in the graph network, which is much greater than the range of 55.7–87.07 per cent reported in previous literature. Thus, it has been concluded that the GCN with node classification models has improved the accuracy.

When accessing structured contents in XML form, information requests are formulated in the form of special query languages such as NEXI, Xquery, etc. However, it is not easy for end users to compose such information requests using these special queries because of their complexities. Hence, the purpose of this paper is to automate the construction of such queries from common query like keywords or form-based queries.

In this paper, the authors address the problem of constructing queries for XML retrieval by proposing a semantic-syntax query model that can be used to construct different types of structured queries. First, a generic query structure known as semantic query structure is designed to store query contents given by user. Then, generation of a target language is carried out by mapping the contents in semantic query structure to query syntax templates stored in knowledge base.

Evaluations were carried out based on how well information needs are captured and transformed into a target query language. In summary, the proposed model is able to express information needs specified using query like NEXI. Xquery records a lower percentage because of its language complexity. The authors also achieve satisfactory query construction rate with an example-based method, i.e. 86 per cent (for NEXI IMDB topics) and 87 per cent (NEXI Wiki topics), respectively, compare to benchmark of 78 per cent by Sumita and Iida in language translation.

The proposed semantic-syntax query model allows flexibility of accommodating new query language by separating the semantic of query from its syntax.

The purpose of this paper is to propose and evaluate XKQSS, a query structuring method that relegates the task of generating structured queries from a user to a search engine while retaining the simple keyword search query interface. A more effective way for searching XML database is to use structured queries. However, using query languages to express queries prove to be difficult for most users since this requires learning a query language and knowledge of the underlying data schema. On the other hand, the success of Web search engines has made many users to be familiar with keyword search and, therefore, they prefer to use a keyword search query interface to search XML data.

Existing query structuring approaches require users to provide structural hints in their input keyword queries even though their interface is keyword base. Other problems with existing systems include their inability to put keyword query ambiguities into consideration during query structuring and how to select the best generated structure query that best represents a given keyword query. To address these problems, this study allows users to submit a schema independent keyword query, use named entity recognition (NER) to categorize query keywords to resolve query ambiguities and compute semantic information for a node from its data content. Algorithms were proposed that find user search intentions and convert the intentions into a set of ranked structured queries.

Experiments with Sigmod and IMDB datasets were conducted to evaluate the effectiveness of the method. The experimental result shows that the XKQSS is about 20 per cent more effective than XReal in terms of return nodes identification, a state-of-art systems for XML retrieval.

Existing systems do not take keyword query ambiguities into account. XKSS consists of two guidelines based on NER that help to resolve these ambiguities before converting the submitted query. It also include a ranking function computes a score for each generated query by using both semantic information and data statistic, as opposed to data statistic only approach used by the existing approaches.