Enhancing Learning Object Recommendations for Teachers Using Adaptive Neighbor Selection

Towards comprehensive approaches for the rating prediction phase in memory-based collaborative filtering recommender systems

Educational Recommender Systems aim to provide students with search relevant results adapted to their needs or preferences and delivering those educational contents such as Learning Objects (LOs) that could be closer than expected. LOs can be defined as a digital entity involving educational design characteristics. Each LO can be used, reused, or referenced during computer-supported learning processes, aiming at generating knowledge, skills, attitudes, and competences based on the student profile. The aim of this paper is to present a student-centered LO recommender system based on a hybrid recommendation technique that combines three following approaches: content-based, collaborative and knowledge-based. In addition, those LOs adapted to the student profile are retrieved from LO repositories using the stored descriptive metadata of these objects. A testing phase with a case study is performed in order to validate the proposed hybrid recommender system that demonstrates the effectiveness of using this kind of approaches in virtual learning environments.KeywordsStudent-centered hybrid recommender systemsLearning objectsMetadataRepositories

Collaborative filtering (CF), the most successful information filtering technique for recommender systems, is either memory-based or model-based. While the former is more accurate, its scalability compared to model-based is poor. Moreover, the similarity functions used by most recommender systems are compensatory and allow very high (pros) and very low (cons) scores to compensate each other. This paper presents a hybrid movie recommender system that retains memory-based CF accuracy, model-based CF scalability, and alleviates the compensation problem of similarity functions. The proposed recommender system relies on a compact user model and fuzzy concordance/discordance principle. The user model speeds up the online process of generating a set of like-minded users within which a memory-based CF is carried out. The inter users comparison is done by using fuzzy concordance/discordance principle to alleviate the similarity compensation problem. The pros and cons between users are measured separately and then the overall statement about them is obtained by balancing the pros and cons within the set of criteria. Besides our approach is fast and compact, computational results reveal that it outperforms the classical one.

Recommender systems aim to help users to find items that fit their requirements and preferences. In that field, the collaborative filtering (CF) approach is considered as a widely used one. There are two main approaches for CF: memory-based and model-based. Both of the two approaches are based on the use of users' ratings to predict the top-N recommendation for the active user. Despite its simplicity and efficiency, The CF approach stills suffer from many drawbacks including sparsity, gray sheep and scalability. The aim of this work is to deal with the gray sheep problem, by proposing a novel collaborative filtering approach. This novel approach aims to enhance the accuracy of prediction by turning the users whose preferences disagree with the target user, into new similar neighbors. For instance, if a user X is dissimilar to a user Y then the user $\lnot$X is similar to the user Y.To evaluate the performance of the proposed approach, we have used two datasets including MovieLens and FilmTrust. The Experimental results show that our approach outperforms many traditional recommendation techniques.

This paper discusses some important issues regarding the the management of Learning objects covering searching over repositories and different approaches of recommendation systems and presents a multiagent system based application model for indexing, retrieving and recommending learning objects stored in different and heterogeneous repositories. The objects within these repositories are described by filled fields using different metadata (data about data) standards. The searching mechanism covers several different learning object repositories and the same object can be described in these repositories by the use of different types of fields. Aiming to improve accuracy and coverage in terms of recovering a learning object and improve the relevance of the results we propose an information retrieval model based on a multiagent system approach and an ontological model to describe the covered knowledge domain.

A recommendation system is a method that provides suggestions of items that might users like. There are many domains that can be recommended, one of the most demanded domains by users today is food. In the era of big data, food choices from the large amount of data make it difficult for users to choose the right food for them. The collaborative filtering (CF) approach is considered capable of providing accurate and high quality item suggestions. One of the algorithms that can provide good performance results from the CF approach is Matrix Factorization (MF). This study aims to test a dataset that contains product ratings of food items using three MF algorithms, which are Singular Value Decomposition (SVD), SVD with Implicit Ratings (SVD++), and Non-Negative Matrix Factorization (NMF). Different latent factors are also used for the purpose of improving the performance of the proposed recommendation system algorithm. The dataset used is Amazon Fine Food Reviews. The study shows NMF and SVD++ as the best algorithm for generating user rating predictions for items. NMF has the smallest average prediction error as measured by MAE which is 0.7311. While SVD++ obtains the smallest prediction error value of 1.0607 as measured using RMSE. In addition to these results, the top-n evaluation also shows that the proposed algorithm performs quite well. The hit ratio value for each different n-item always increases proportionally to the number of recommended n-items. The highest hit ratio value is generated from the SVD++ algorithm of 0.0025 on n-item recommendations of 25 items. Overall it can be said that the proposed algorithm has good performance in providing item recommendations.

Ubiquitous Learning (U-Learning), as an emerging learning paradigm, makes it possible for learners to carry out the learning activities at any places and at anytime. With the advantages of the devices, learners can obtain a variety of supplementary materials from the internet. In the scope of distance learning, LOR (Learning Object Repository) stands for managing and sharing of learning related materials (known as learning objects). However, some challenges may raise while performing these activities. For instance, a huge amount of learning objects may appear while learners utilize the search service provided by LOR. Learners have to spend time on collecting relevant resources for specific purposes. This situation may discourage the reusability of learning objects especially in a ubiquitous environment. In this paper, based on systematic re-examination of reuse scenarios, an adaptive service, as a resource discovery and search middleware, was proposed to assist learners in obtaining possible objects under ubiquitous environment. The achievement of proposed mechanism can reorganize the search results in order of similarity degree based on the mixed information including the usage history, geographical information and query criteria. It may make resource retrieval more efficient in u-learning environment. As another contribution, the emerging ubiquitous device, Apple iPad/iPod, was utilized to be the major end-user testbed.

Learning Objects (LOs) have emerged as a cornerstone approach for the development and distribution of educational content. These resources are distributed by Learning Object Repositories (LORs), which can make it easier for users to find suitable LOs by using Recommender Systems (RSs). This paper presents a hybrid recommendation model for LORs that combines content-based, demographic and context-aware techniques, along with the use of quality and popularity metrics. This article also describes how the model has been used to implement two RSs for two real LORs: ViSH and Europeana. Each of these RSs was evaluated in terms of accuracy, utility,usability and satisfaction perceived by end users. Besides, an A/B testing was performed in ViSH to compare the recommendations of the RS with random suggestions. The results showed that the RSs had a high user acceptance in terms of utility, usability and satisfaction, and that the RSs significantly exceeded the performance achieved by the random recommendations.

Collaborative filtering and deep learning based recommendation system for cold start items

The Recommendation System plays an increasingly important role in our daily lives. With the increasing amount of information on the internet, the recommendation system can also solve problems caused by increasing information quickly. Collaborative filtering is one method in the recommendation system that makes recommendations by analyzing correlations between users. Collaborative filtering accumulates customer item ratings, identifies customers with common ratings, and offers recommendations based on inter-customer comparisons. This study aims to build a system that can provide recommendations to users who want to order or choose fast food menus. This recommendation system provides recommendations based on item data calculations with customer review data using a collaborative filtering approach. The results of applying cosine similarity calculation to determine fast food menu recommendations obtained for the item-based recommendation is Pizza Frankfurter BBQ Large with a value of 1.0, item-based with genre recommendation is Calblend Float with value 1.0 and user-based recommendation is Pizza Black Pepper Beef / Chicken Large with mean score 2.5.

Recommender system is the system which, by using automated information filtering technology, recommends products or services to the customers who are likely to be interested in. Those systems are widely used in many different Web retailers such as Amazon.com, Netfix.com, and CDNow.com. Various recommender systems have been developed. Among them, Collaborative Filtering (CF) has been known as the most successful and commonly used approach. CF identifies customers whose tastes are similar to those of a given customer, and recommends items those customers have liked in the past. Numerous CF algorithms have been developed to increase the performance of recommender systems. However, the relative performances of CF algorithms are known to be domain and data dependent. It is very time-consuming and expensive to implement and launce a CF recommender system, and also the system unsuited for the given domain provides customers with poor quality recommendations that make them easily annoyed. Therefore, predicting in advance whether the performance of CF recommender system is acceptable or not is practically important and needed. In this study, we propose a decision making guideline which helps decide whether CF is adoptable for a given application with certain transaction data characteristics. Several previous studies reported that sparsity, gray sheep, cold-start, coverage, and serendipity could affect the performance of CF, but the theoretical and empirical justification of such factors is lacking. Recently there are many studies paying attention to Social Network Analysis (SNA) as a method to analyze social relationships among people. SNA is a method to measure and visualize the linkage structure and status focusing on interaction among objects within communication group. CF analyzes the similarity among previous ratings or purchases of each customer, finds the relationships among the customers who have similarities, and then uses the relationships for recommendations. Thus CF can be modeled as a social network in which customers are nodes and purchase relationships between customers are links. Under the assumption that SNA could facilitate an exploration of the topological properties of the network structure that are implicit in transaction data for CF recommendations, we focus on density, clustering coefficient, and centralization which are ones of the most commonly used measures to capture topological properties of the social network structure. While network density, expressed as a proportion of the maximum possible number of links, captures the density of the whole network, the clustering coefficient captures the degree to which the overall network contains localized pockets of dense connectivity. Centralization reflects the extent to which connections are concentrated in a small number of nodes rather than distributed equally among all nodes. We explore how these SNA measures affect the performance of CF performance and how they interact to each other. Our experiments used sales transaction data from H department store, one of the well?known department stores in Korea. Total 396 data set were sampled to construct various types of social networks. The dependant variable measuring process consists of three steps; analysis of customer similarities, construction of a social network, and analysis of social network patterns. We used UCINET 6.0 for SNA. The experiments conducted the 3-way ANOVA which employs three SNA measures as dependant variables, and the recommendation accuracy measured by F1-measure as an independent variable. The experiments report that 1) each of three SNA measures affects the recommendation accuracy, 2) the density's effect to the performance overrides those of clustering coefficient and centralization (i.e., CF adoption is not a good decision if the density is low), and 3) however though the density is low, the performance of CF is comparatively good when the clustering coefficient is low. We expect that these experiment results help firms decide whether CF recommender system is adoptable for their business domain with certain transaction data characteristics.

The Indian market for internet shopping has already been penetrated by E-Commerce. People are increasingly attracted to shop online, which has altered the market landscape. Amazon, Snap deal, Flipkart, and more companies offer online shopping sites that are gaining traction. As the number of online buyers and dealers grows, smart business approaches are required to deal with the massive amounts of data generated every day. Recommendation systems play a crucial role in sifting this data and giving consumers with relevant information. Various strategies for recommender systems have been used, such as Collaborative Filtering, Content-based, and Demographic, but these techniques have a number of flaws that make them ineffective. Among these collaborative filtering approach is widely popular because of its simplicity and accuracy. But still these systems can be improved in their accuracy when used for multiple domains. As a result, more distinguishing traits must be identified in order to optimize these procedures. This can be accomplished by combining the benefits of multiple strategies in a hybrid approach. This paper describes a hybrid approach that identifies the users with shared preferences (semantic users) and their dynamic interests with parallel training to improve recommendation efficiency. Results were compared with other state of art techniques and the accuracy of our system is found to be improved by 2% and seems to give promising results for multiple domains.

We present an approach for the personalized access and the students’ coauthoring in repositories of Learning Objects (LOs). The personalized access combines content-based and collaborative filtering mechanisms. Students’ coauthoring includes the incorporation of assessments of the existing LOs and new LOs which are peer reviewed.

The increasing demand for World Wide Web (Web) based courses has imposed some challenges in Information and Communications Technology (ICT) field, in particular when considering the education supported by the Web. One of the emerging needs of education supported by the Web is related to the Learning Objects (LO) usage inside Learning Management Systems (LMS). Besides production of "good quality" LO, it must be provided efficient methods concerning their storage and dissemination. However, some difficulties have been reported by teachers of Web based courses: "expensive costs" for LO authoring; lack of tools for sharing these materials; and negative practical experiences for reusing them in an LMS. Some proposals to mitigate these problems have been the development of Learning Object Repository (LOR) tools, mainly in order to facilitate the dissemination of LO. Despite the progress already achieved, there are still several barriers for their adoption by teachers such as lack of easy integration between LMS environment and LOR. In general, current integrations are not simple for typical users. In this context, the target of this project is to presents an innovative LOR, the Interactive Repository of Digital Contents -iRepository. Its model provides seamless integration with a particular LMS and offers several resources to help teaching and learning processes, such as registry of LO educational effectiveness. As a practical contribution, the fundamentals of iRepository model were implemented for an specific LMS, the Moodle system.

Collaborative filtering (CF) is the most popular approach in personalized recommender systems. Although CF approaches have successfully been used and have the advantage in that it is unnecessary to analyze item content when generating recommendations, they nevertheless suffer from problems with accuracy. In this paper, we propose a new CF approach to improve recommendation performance. First, a new information entropy-driven user similarity measure model is proposed to measure the relative difference between ratings. A Manhattan distance-based model is then developed to address the fat tail problem by estimating the alternative active user average rating. The effectiveness of the proposed approach is analyzed on public and private data sets. As a result of the introduction of the new similarity measure and average rating estimation, we demonstrate that the proposed new CF recommendation approach provides better recommendations.