Extracting feature requests from online reviews of travel industry

• Autor(a) principal: Kumari, Superna
• Data de Publicação: 2022
• Outros Autores: Memon, Zulfiqar Ali
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Acta scientiarum. Technology (Online)
• Texto Completo: http://www.periodicos.uem.br/ojs/index.php/ActaSciTechnol/article/view/58658
• topic: Requirement engineering; feature requests; user reviews; app stores; travel booking applications; sentiment analysis; Requirement engineering; feature requests; user reviews; app stores; travel booking applications; sentiment analysis

Before product development, Requirement Engineering (RE) is the fundamental need to know customer preferences for any product. Traditionally, RE is carried out in several ways, particularly by conducting interviews, questionnaires, surveys etc. but these methods provide limited amount of data. As user’s preferences vary from country to country for any type of application, it is very hectic and time consuming to collect user requirements from different countries manually. As the internet is widely used now a days, a large number of customer’s reviews are available online that can be used to obtain the requirements for any product without manual work. Online customer reviews can be divided into three types: user experience, bugs and feature requests. Among these 3 categories, feature requests can be very useful for stakeholders (analysts/ requirements engineers) to acquire the requirements of each application. So, the approach is proposed for feature requests extraction from mobile application reviews of travel industry. In this paper, 4 categories of mobile apps of travel industry belonging to 5 countries have been extracted from Google Play Store and Apple Store. For each category, data from 5 different mobile applications have been considered. Since, the review of users from different countries is in their respective language, those reviews are translated into a standard language i.e. English, and then feature requests from these reviews have been extracted. After that, features are retrieved from user reviews and topic modeling is performed on extracted features since one or more features can be modelled under one topic. To know the opinions of users for any feature request, sentiment analysis is also performed on feature request sentences. These feature requests are also classified as Functional and Non-functional Requirements since it is very useful for application developers to improve or maintain the product to better facilitate the application users