User-opinion mining for mobile library apps in China: exploring user improvement needs

To further provide some insight into mobile library (m-library) applications (apps) user needs and help libraries or app providers improve the service quality, the purpose of this paper is to explore all the types of user improvement needs and to discover which need is the most important based on user results.

Data were collected from more than 27,000 m-library app users from 16 provinces and autonomous regions in China. Text analysis using latent Dirichlet allocation and Word2Vec was carried out by text preprocessing. Furthermore, a visual presentation was conducted through pyLDAvis and word cloud. Finally, combined with expert opinions, the results were summarized to find the different types of needs.

There are three different types of needs for improvement: needs of function, needs of technology and needs of experience. These types can be further divided into six subtypes: richness of function, feasibility of function, easiness of technology, stableness of technology, optimization of experience and customization of experience. Besides the richness of function, the feasibility of function has received the most attention from users.

Most studies on m-library user needs have only focused on a method of quantitative research based on questionnaire surveys. This study, however, is the first to apply text mining methods for large-scale user opinion texts, which place more focus on user needs and inspire libraries and app providers to further improve their services.