Search results

Airbnb Experiences is a new type of service launched by Airbnb in November 2016, where users can offer travellers a wide range of activities. This study devotes attention to analysing customer feedback expressed in online reviews published in Airbnb to evaluate those experiences.

A total of 1,110 reviews were collected from 12 categories, including 111 experiences, resulting in 10 reviews per experience. First, the sentiment score was computed based on the text of the reviews. Second, 17 quantitative features encompassing user, Airbnb experience and review information were used to model the score through a support vector machine. Third, a sensitivity analysis was performed to extract knowledge on the most relevant features influencing the sentiment score.

Tourists writing online reviews are not only influenced by their tourist experience but also by their own online experience with the booking and online review platform. The number of reviews made by the user accounted for more than 20 per cent of relevance, while users with more reviews tended to grant more positive reviews.

Current literature is enhanced with a conceptual model grounded on existing studies that assess tourist satisfaction with tour services. Both services online visibility and user characteristics have shown significant importance to tourist satisfaction, adding to the existing body of knowledge.

This paper aims to propose a data mining approach to evaluate a conceptual model in tourism, encompassing a large data set characterized by dimensions grounded on existing literature.

The approach is tested using a guest satisfaction model encompassing nine dimensions. A large data set of 84 k online reviews and 31 features was collected from TripAdvisor. The review score granted was considered a proxy of guest satisfaction and was defined as the target feature to model. A sequence of data understanding and preparation tasks led to a tuned set of 60k reviews and 29 input features which were used for training the data mining model. Finally, the data-based sensitivity analysis was adopted to understand which dimensions most influence guest satisfaction.

Previous user’s experience with the online platform, individual preferences, and hotel prestige were the most relevant dimensions concerning guests’ satisfaction. On the opposite, homogeneous characteristics among the Las Vegas hotels such as the hotel size were found of little relevance to satisfaction.

This study intends to set a baseline for an easier adoption of data mining to evaluate conceptual models through a scalable approach, helping to bridge between theory and practice, especially relevant when dealing with Big Data sources such as the social media. Thus, the steps undertaken during the study are detailed to facilitate replication to other models.

This study aims to propose a model to explain online review helpfulness grounded on both previously identified constructs (e.g. review length) and new ones, which have been analyzed in other online reviews’ contexts but not to explain helpfulness.

A total of 112,856 reviews published in TripAdvisor about 21 Las Vegas hotels were collected and a random forest model was trained to assess if a review has received a helpful vote or not.

After confirming the validity of the proposed model, each of the constructs was evaluated to assess its contribution to explaining helpfulness. Specifically, a newly proposed construct, the response lag of the manager’s replies to reviews, was among the most relevant constructs.

The achieved results suggest that hoteliers should invest not only in responding to the most interesting reviews from the hotel’s perspective but also that they should do it quickly to increase the likeliness of the review being considered helpful to others.

论：项解释酒店业中在线评论有用性的模型

研究目的：本论文提出一项模型, 用于解释在线网络的有用性, 这项模型基于之前文献定义的结构（如评论长短）以及在其他在线评论题材下分析的新结构。而这些结构之前未用来解释有用性。

研究设计/方法/途径：样本数据为112,856条TripAdvisor关于21家拉斯维加斯酒店的评论。本论文创建了一种随机森林模型, 用于检测是否一条评论是有用的。

研究结果：本论文肯定了提出模型的有效性, 此外, 本论文评估了每个模型结构, 检测其是否解释了有用性。具体地说, 本论文提出了一种新的结构, 经理回复评论的时间长短, 是最相关的结构之一。

研究原创性/价值：研究结果表明, 酒店经理应该不仅从酒店角度出发回复最有意思的评论, 而且要快速回复, 以增加其他用户评价这条评论是否对他人决策的有用性。

Virtual reality (VR) and augmented reality (AR) are two technological breakthroughs that stimulate reality perception. Both have been applied in tourism contexts to improve tourists’ experience. This paper aims to frame both AR and VR developments during the past 15 years from a scientific perspective.

This study adopts a text mining and topic modelling approach to analyse a total of 1,049 articles for VR and 406 for AR. The papers were selected from Scopus, with the title, abstract and keywords being extracted for the analysis. Formulated research hypotheses based on relevant publications are then evaluated to assess the current state of the broader scope of the large sets of literature.

Most of research using AR is based on mobile technology. Yet, wearable devices still show few publications, a gap that is expected to close in the near future. There is a lack of research adopting Big Data/machine learning approaches based on secondary data.

As both AR and VR technologies are becoming more mature, more applications to tourism emerge. Scholars need to keep pace and fill in the research gaps on both domains to move research forward.

论旅游业中AR和VR的近期发展

摘要

研究目的 – VR和AR是两种提高人们现实感知的科技突破。二者均被运用到旅游场景中提高游客体验。本论文旨在建立一个模型来描述过去15年内AR和VR的发展轨迹。

本论文采用文本挖掘和主题建模方法来分析1049篇有关VR和406篇有关AR的文章。样本库采样于Scopus, 通过题目、摘要、和关键词来检索。根据检索结果, 本论文提出研究假设, 并且审视延展更大范围的相关文献。

本论文发现大多数AR文章都是以移动技术为本。然而, 可穿戴设备始终有很少的文章。因此, 文献缺口预计在不久的将来会得以补上。研究还发现, 基于二手数据而运用大数据/机器学习方法的文章少之又少。

随着AR和VR技术越来越成熟, 旅游业更多的运用案例随之出现。学者需要紧随脚步, 填补AR和VR的文献缺口, 推进研究进程。

关键词：虚拟现实、增强现实、文献分析、旅游业