تحليل المشاعر القائم على الذكاء الاصطناعي كأداة لتقييم فعالية خدمات المعلومات

الملخص EN

The study explores the opportunities of using sentiment analysis technique, as one of the artificial intelligence applications, to measure the user satisfaction on the library and information services as an indicator for its effectiveness.

The study reviews main tools and techniques used, the areas of application in Arabic libraries’ environment through the analysis of users’ comments published in its social media pages.

The study discusses the challenges of implementing this method in the Arabic libraries’ environment in terms of the significance of the comments, the polarity classification process to positive, negative, or neutral feelings about the information services provided which supports the Decision-making process to improve the library and information services provided.

An exploratory study was conducted for a sample of users’ comments on social media pages.

A dataset of (665) tweets/comments was collected through December 2021 to end of April 2022, by searching the terms related to library services via Twitter and Facebook platforms.

The NLP process was performed using the free tool Google Colab.

The result of the analysis shows the capability of this method to provide a detailed data with required analysis in various forms which help identify the user opinions towards the library posts/tweets.

There are some sentiment analysis applications that can be used to generate the sentiments automatically, but they provide limited capabilities especially with Arabic text.

Some challenges are identified related to the nature of Arabic language, and its dialects in the NLP process, and generating the word cloud, the availability of software tools and code libraries to deal with Arabic text effectively.

Some other challenges are related to the nature of the comments/tweets and their meaning ambiguity, using Franco-Arab to express the Arabic text, the lack of active participation of users in the terms of their comments, in addition to the challenges related to privacy issues.

The study recommends the collaboration among the researchers from disciplines of library and information sciences, computer science, and linguistics to conduct interdisciplinary studies to design ready-made software tools that support the capabilities of the Arabic language and can be used by libraries and other information institutions directly without the need for the programming complexities to get detailed opinion analysis data.