Sentiment analysis of hotel reviews
Download 0.86 Mb. View original pdf
|
| 3 | P age b INTRODUCTION TripAdvisor is a solution to the travelers who seek to collect information about a destination. One point of concern for travelers have been a consolidated forum that gives them reviews about destinations and place of lodgings from fellow travelers without any bias. As the trend towards travelling is increasing these days, people tend to read and research about the places they want to travel before actually going to that place. TripAdvisor gives a platform for people to search about accommodation and tours mainly through use of reviews. Rating is becoming increasingly important for customers, which helps them obtain useful information of places more easily and more efficiently. While a rating can be a good indicator for the opinion, text reviews are usually more elaborative. At the same time, reading text reviews is also time-consuming. It is of strategic importance to be able to extract the crux of a review without reading multiple lines of text. This process of computationally identifying and categorizing opinions expressed in apiece of text, especially in order to determine whether the writer's attitude towards a particular topic, product, etc. is positive, negative, or neutral is called sentiment analysis. The result of this is not only useful to users to see which places are of superior quality but also to Hoteliers, who would then be able to track their service reception through customer feedbacks. This classification can also be done through analysis of individual aspects involved in a hospitality service say price point, service, etc. The results of aspect-based semantic analysis can be used in a wide variety of applications, including building better-customized recommendation systems and catering to the right segment in accordance with one’s own serving capabilities. In our analysis we have used data used from The database and Information Systems Laboratory The data consists of 42,431 many textual reviews (rows) of around 15 hotels. Along with the reviews there are ratings of various aspects namely service, cleanliness, internet access, reception, value, sleep quality, location and rooms. Additionally, an overall rating is also present. All these ratings are on a point uniform scale, 5 standing for excellent service and span over the time period of While it’s difficult to speculate how a relatively immature system might evolve in the the future, there is a general assumption that sentiment analysis needs to move beyond a one-dimensional positive to negative scale. We have used a three-dimensional sentiment analysis classifying reviews as bad, neutral or good. For the future, to truly understand and capture the broad range of emotions that humans express as written word, we need a more sophisticated multidimensional scale. |