Towards a Precision-Oriented Multi-Lingual Semantics-Based Cultural Heritage Recommender System رسالة ماجستير

Abstract

The cultural heritage sector and its associated tourism facilities have been affected notably by the advancement of the Internet, as well as the explosive growth of smartphones and other handheld devices. These days, visitors have access to reliable and trusted content related to cultural heritage sites world-wide. They can access this information either using the Internet (via Web interfaces) or by using their handheld devices. Considering the latter approach, a user can search for any desired site, and details that pertain to it will be presented in a timely fashion and neat interface. However, conventional cultural heritage information systems lack the ability to adapt their behavior to the preferences, needs, interests and other features that are required by users (be they single tourists or tourism groups). In this research work, we address the issue of designing an effective multi-lingual semantics-based mobile recommender system about Palestine’s cultural heritage. We facilitated users’ access to cultural heritage content by providing them with multiple search functionalities. In this context, a user can search for cultural heritage sites or topics via a dedicated interface; wherein the system takes a given query as input and retrieves all relevant cultural heritage documents based on their semantic similarity. Accordingly, users can express their information needs using keywords (a.k.a. tags) or sentence-like queries to describe their information needs. The proposed system processes users’ queries by utilizing natural language processing techniques, multi-lingual ontologies and other term relatedness measures. A second option is to search by using current location (to retrieve historical places and events associated with the place where the visitor is located), considering network availability. A third option allows users to capture and submit images as input and the system accordingly retrieves all relevant results based on their content similarity, such as texts or objects detected and recognized in the given images, where users can then share the returned result with their friends. Moreover, the proposed VIII mobile-based application is aimed to adapt itself to the user preferences and information needs; for personalizing their experience on one hand, and offering more effective and efficient interaction on the other. In this context, the system acquires user queries and automatically and gradually deliver its output to meet the user’s information needs and preferences based on her/his logging information. To evaluate the quality of the utilized techniques, we have developed a prototype of the proposed mobile-based application and tested it using Android devices and a manually-constructed ontology (henceforth named as Holy-Land Ontology) that we have enriched with links to the Art & Architecture Thesaurus (AAT) and DBpedia semantic thesaurus about cultural heritage information. When we compare our system with other systems in this field, findings demonstrated that our system provides additional search features and functionalities to users. The implementation of machine learning techniques to extract and recognize objects from images has also helped users to better understand the content of the images that they captured. In addition, our constructed ontology is the first one that address cultural heritage in the holy-land region.