Date of Award
Master of Science (MS)
Computer Science and Engineering
Krenare Pireva Nuqi
Avoiding digital marketing, surveys, reviews and online users behavior approaches on digital age are the key elements for a powerful businesses to fail, there are some systems that should preceded some artificial intelligence techniques. In this direction, the use of data mining for recommending relevant items as a new state of the art technique is increasing user satisfaction as well as the business revenues. And other related information gathering approaches in order to our systems thing and acts like humans. To do so there is a Recommender System that will be elaborated in this thesis. How people interact, how to calculate accurately and identify what people like or dislike based on their online previous behaviors. The thesis includes also the methodologies recommender system uses, how math equations helps Recommender Systems to calculate user’s behavior and similarities. The filters are important on Recommender System, explaining if similar users like the same product or item, which is the probability of neighbor user to like also. Here comes collaborative filters, neighborhood filters, hybrid recommender system with the use of various algorithms the Recommender Systems has the ability to predict whether a particular user would prefer an item or not, based on the user’s profile and their activities. The use of Recommender Systems are beneficial to both service providers and users. Thesis cover also the strength and weaknesses of Recommender Systems and how involving Ontology can improve it. Ontology-based methods can be used to reduce problems that content-based recommender systems are known to suffer from. Based on Kosovar’s GDP and youngsters job perspectives are desirable for improvements, the demand is greater than the offer. I thought of building an intelligence system that will be making easier for Kosovars to find the appropriate job that suits their profile, skills, knowledge, character and locations. And that system is called TROI Search engine that indexes and merge all local operating job seeking websites in one platform with intelligence features. Thesis will present the design, implementation, testing and evaluation of a TROI search engine. Testing is done by getting user experiments while using running environment of TROI search engine. Results show that the functionality of the recommender system is satisfactory and helpful.
Mulaj, Donjeta, "THE USE OF RECOMMENDER SYSTEMS IN WEB APPLICATIONS – THE TROI CASE" (2018). Theses and Dissertations. 3.