Symposia & Conferences
Permanent URI for this communityhttp://repository.kln.ac.lk/handle/123456789/10213
Browse
3 results
Search Results
Item Text mining approach on on-line guest reviews: A case study from hospitality industry.(International Research Symposium on Pure and Applied Sciences, 2017 Faculty of Science, University of Kelaniya, Sri Lanka., 2017) Hewage, N. I.; Jayalal, S.Social media and consumer-generated context on the Internet have become an integral part of the modern society with millions of users. With the increase of its popularity, tourism industry has shifted towards electronic transactions. Tourists now tend to use online social media reviews and ratings posted by hotel guests to make decisions before booking a hotel, which is an impossible task for a single user due to its high volume of reviews. Hence, to make better decisions, ranking of hotels for a specific region will be beneficial for the tourists who are willing to travel in that region and for the management of the hotel as well. However, while a handful of studies have employed on hotel guest satisfaction and experience by analyzing online hotel guest reviews collected from online travel agencies, there is a significant research problem with ranking hotels by analyzing hotel guest reviews in aspect level consideration. Expedia.com, Agoda.com and Booking.com are some of the leading online travel agencies that have millions of users. Online reviews used for this study are collected from 18 hotels that belong to all these three online travel agencies, and from that dataset, 6 hotels are selected as testing dataset. The dataset contains reviews from year 2010 to 2016. In this study, we propose a ranking mechanism, that ranks hotels by using the overall rating values, sentiment scores and the reviewed year. For computing sentiment scores, each review is split into sentences and they are categorized in to six attributes as Location, Service quality, Cleanliness, Comfort of rooms, Value for money and other. Thereafter, the sentiment analysis is done by considering the weight of the positive and negative words. In this research, we present a novel ranking algorithm to rank hotels, considering the reviewed year and computing the ranking score by getting the variance of the polarity rate and variance of rate of overall rating from initial year to the last. The results were taken by considering specific time period and without considering specific time period. Therefore, when using all the reviews without considering a time period, the rankings deviate from the Booking.com and TripAdvisor.com rankings. When using the reviews within 3 years of time, the ranking results are almost equal to the TripAdvisor.com rankings. When the time period is reduced for 18 months, the accuracy is 50% with Booking.com rankings and 33.3% with TripAdvisor.com rankings. This could be due to the fact that the above mentioned online travel agencies use online reviews more than two years, and therefore it perhaps causes for the deviation of rankings. Since sample dataset is used in this study, the accuracy can be increased by using a large dataset. Since this ranking mechanism considered variance to clarify the performance of hotels, and it is not only depending on the number of positive reviews or star ratings of the hotel, this is beneficial for the hotels which are not much popular, but having good standards.Item Detection of cyber bullying on social media networks(Faculty of Science, University of Kelaniya, Sri Lanka, 2016) Priyangika. S.; Jayalal, S.Social Media is becoming an integral part of people’s daily lives today. It is an effective way of sharing one’s life experiences, special occasions, achievements and other events with their friends and family. Although it is a fruitful way to communicate with groups, some people find themselves being insulted or offended by others who are involved in certain post or conversations. These insulations can be based on racism, using profanity or any other vulgar or lewd language. This cyber bullying needs to be monitored and controlled by the social media site owners since it will highly effect on the number and safety of the active site membership. Currently, there is no automated process of identifying offensive comments by the social network site itself. It can be only diagnosed by humans after reading the comments, flagging or reporting them to the owner of the site or blocking the offender. Considering the massive big data set generated in social media daily, automatically detection of offensive statements is required to reduce insulation effectively. For this purpose, text classification approach can be applied where a given text will be categorized as insulting or not, through learning from a pre-learned model. In order to develop the model, data was collected from the popular data repository site named www.kaggle.com. The dataset consists of comments posted on Facebook and Twitter. Firstly the dataset was divided into training data set and test data set. Then the collected data was preprocessed by removing the unwanted strings, correcting words and eliminating duplicate data fields. In the next step, features or keywords were extracted which are qualified to distinguish a statement as ‘insulting’ using N-grams model and counting methods. Feature selection is done using Chi- Squared test and finally apply classification algorithms for separating insulting comments and non-insulting comments from a dataset given. Machine learning algorithms such as Support Vector Machines (SVM), Naïve Bayes, Logistic Regression and Random Forest are used for this. Out of the classification algorithms, SVM is to be performed better than other algorithms since this is a two-class classification problem and a comment is to be classified only into two separate classes which are ‘insulting’ and ‘neutral’. With an exact separation of a given comment into ‘insulting’ and ‘neutral’ category, cyberbullying happening through offensive comments posted on social media sites can be detected.Item Predicting box office success of movies using sentiment analysis and opinion mining(Faculty of Science, University of Kelaniya, Sri Lanka, 2016) Basnayake, H.; Jayalal, S.Movies and social media come together as a result of people sharing their opinions on social media and movie makers using the same platforms for movie promotions. From movie makers to movie goers, many parties are interested in the success or failure of a movie. Forecasting the success of a movie before its release has been a difficult task for many industry analysts. Since film industry’s unpredictable nature, many analysts have come up with different algorithms and mechanisms to predict the success of a movie. One of the mechanisms to predict the box office success is hype analysis. Hype is one of the factors that drive people to the theatres to watch a new movie. Box office opening of a new movie depends on this hype and it will boost up the total box office collection. Hype can be estimated through social media platforms like Twitter. Twitter can be used as a corpus for sentiment analysis and opinion mining. A movie’s success cannot be predicted in a high accurate level solely based on social factors. Classical factors like movie’s brand name, cast, director, etc. are also important aspects in movie’s performance at box office and should be considered as well. However, a highly accurate method for movie box office prediction integrating both social and classical factors is yet to be introduced for this research area. In this study, tweets related to the particular movie before releasing are collected using an archiver tool and are used as input data. Then the collected data is preprocessed in order to get a clean dataset. As a part of sentiment analysis and opinion mining, feature selection is performed using N-gram method in order to filter out irrelevant data records and unlike Bag of words method, this does not require an extensive dictionary of words since it uses combinations of words and letters. Afterwards the data related to classical factors are integrated with the proposed formula in order to predict the opening box office collection of the movie. The proposed formula is an extension of a formula used in a previous research and the new extension represent the inclusion of classical factors. Finally, the results are compared with actual box office data and the previous formula results in order to compare and determine the level of accuracy. Based on initial results, the proposed formula showed of an accuracy level more than 85 percent when the results were compared with actual box office data. Even though it produced a higher accuracy level, the results produced were less than the actual box office values. Thus further testing is needed to determine the actual accuracy level.