الفهرس 
Chapter 1 IntroductionOnly 14 pages are availabe for public view
 |  | from | 119 |  |  |
| | | from | 119 | | |
Abstract
Recommendation Systems (RSs) have gained a great interest in recent days. They record a significant success in several domains including movies, music, news, books, research articles, search queries, and products in general. One of the most important working fields of RS is the Electronic-Learning (E-L) which it can be utilized to overcome many challenges that face hinder users in discovering the most appropriate materials. Several RSs had been introduced which are built on artificial intelligence and soft computing principles. However, they still suffer from either long results or low relevancy. Fog computing technique can enrich E-L based RS as it bridges the gap between the cloud and end devices by enabling computing, storage, networking, and data management on the network nodes within the close vicinity of end devices. Fog computing enables E-Learning communities to expand their services according to their demands. In this thesis , we propose Fog based Recommendation System (FBRS) which can be utilized successfully for promoting the performance of (EL) environment through fog computing .We discuss a framework to consolidate and improve the environment of (EL) through defining three modules (class identification module ”CIM”, subclass identification module ”SIM”, and matchmaking module ”MM”) of FBRS : (i) CIM which calculates the category or class of the desired course according to keyword in user’s query through calculating the relation between query’s concepts and classes’ concepts by using new weighting method and Membership Function techniques. (ii) SIM which calculates the subclass of the desired subject by applying both association rules mining and weighting method based on information gain ratio. Hence, the CIM and SIM are done in the cloud at spaces intervals. (iii) (MM) which retrieves the selected items (courses) and ranks them according to their relevancy to the user’s query by applying Ontology-Based (OB) recommendation and Fuzzy Logic (FL) techniques; this module (MM) actually is done in fog. Moreover, FBRS can employ Fog computing approach to achieve a high response time (e.g., low latency) and security, which is a critical issue for building a good RS.FBRS can overcome many challenges such as personalization and synonymy. Furthermore, it employs several techniques such as association rules, fuzzy logic and ontology so that each technique can solve the defects of the others. Our experiment depends on the Web KB dataset which is the web pages of the computer science department of different universities. The documents were manually classified into four classes; ‘‘Project’’, ‘‘Course’’, ‘‘Faculty’’, and ‘‘Student’’. Experimental results have shown that FBRS outperforms recent techniques in terms of recommendation accuracy.