Knowledge-based recommendation: a review of ontology-based recommender systems for e-learning

Recommender systems in e-learning domain play an important role in assisting the learners to find useful and relevant learning materials that meet their learning needs. Personalized intelligent agents and recommender systems have been widely accepted as solutions towards overcoming information retrieval challenges by learners arising from information overload. Use of ontology for knowledge representation in knowledge-based recommender systems for e-learning has become an interesting research area. In knowledge-based recommendation for e-learning resources, ontology is used to represent knowledge about the learner and learning resources. Although a number of review studies have been carried out in the area of recommender systems, there are still gaps and deficiencies in the comprehensive literature review and survey in the specific area of ontology-based recommendation for e-learning. In this paper, we present a review of literature on ontology-based recommenders for e-learning. First, we analyze and classify the journal papers that were published from 2005 to 2014 in the field of ontology-based recommendation for e-learning. Secondly, we categorize the different recommendation techniques used by ontology-based e-learning recommenders. Thirdly, we categorize the knowledge representation technique, ontology type and ontology representation language used by ontology-based recommender systems, as well as types of learning resources recommended by e-learning recommenders. Lastly, we discuss the future trends of this recommendation approach in the context of e-learning. This study shows that use of ontology for knowledge representation in e-learning recommender systems can improve the quality of recommendations. It was also evident that hybridization of knowledge-based recommendation with other recommendation techniques can enhance the effectiveness of e-learning recommenders.     

An empirical study on the effect of data sparsity and data overlap on cross domain collaborative filtering performance

In the present day, the oversaturation of data has complicated the process of finding information from a data source. Recommender systems aim to alleviate this problem in various domains by actively suggesting selective information to potential users based on their personal preferences. Amongst these approaches, collaborative filtering based recommenders (CF recommenders), which make use of users’ implicit and explicit ratings for items, are widely regarded as the most successful type of recommender system. However, CF recommenders are sensitive to issues caused by data sparsity, where users rate very few items, or items receive very few ratings from users, meaning there is not enough data to give a recommendation. The majority of studies have attempted to solve these issues by focusing on developing new algorithms within a single domain. Recently, cross-domain recommenders that use multiple domain datasets have attracted increasing attention amongst the research community. Cross-domain recommenders assume that users who express their preferences in one domain (called the target domain) will also express their preferences in another domain (called the source domain), and that these additional preferences will improve precision and recall of recommendations to the user. The purpose of this study is to investigate the effects of various data sparsity and data overlap issues on the performance of cross-domain CF recommenders, using various aggregation functions. In this study, several different cross-domain recommenders were created by collecting three datasets from three separate domains of a large Korean fashion company and combining them with different algorithms and different aggregation approaches. The cross-recommenders that used high performance, high overlap domains showed significant improvement of precision and recall of recommendation when the recommendation scores of individual domains were combined using the summation aggregation function. However, the cross-recommenders that used low performance, low overlap domains showed little or no performance improvement in all areas. This result implies that the use of cross-domain recommenders do not guarantee performance improvement, rather that it is necessary to consider relevant factors carefully to achieve performance improvement when using cross-domain recommenders.     

S3Mining: A model-driven engineering approach for supporting novice data miners in selecting suitable classifiers

Data mining has proven to be very useful in order to extract information from data in many different contexts. However, due to the complexity of data mining techniques, it is required the know-how of an expert in this field to select and use them. Actually, adequately applying data mining is out of the reach of novice users which have expertise in their area of work, but lack skills to employ these techniques. In this paper, we use both model-driven engineering and scientific workflow standards and tools in order to develop named S3Mining framework, which supports novice users in the process of selecting the data mining classification algorithm that better fits with their data and goal. To this aim, this selection process uses the past experiences of expert data miners with the application of classification techniques over their own datasets. The contributions of our S3Mining framework are as follows: (i) an approach to create a knowledge base which stores the past experiences of experts users, (ii) a process that provides the expert users with utilities for the construction of classifiers’ recommenders based on the existing knowledge base, (iii) a system that allows novice data miners to use these recommenders for discovering the classifiers that better fit for solving their problem at hand, and (iv) a public implementation of the framework’s workflows. Finally, an experimental evaluation has been conducted to shown the feasibility of our framework.     

Large-scale user modeling with recurrent neural networks for music discovery on multiple time scales

The amount of content on online music streaming platforms is immense, and most users only access a tiny fraction of this content. Recommender systems are the application of choice to open up the collection to these users. Collaborative filtering has the disadvantage that it relies on explicit ratings, which are often unavailable, and generally disregards the temporal nature of music consumption. On the other hand, item co-occurrence algorithms, such as the recently introduced word2vec-based recommenders, are typically left without an effective user representation. In this paper, we present a new approach to model users through recurrent neural networks by sequentially processing consumed items, represented by any type of embeddings and other context features. This way we obtain semantically rich user representations, which capture a user’s musical taste over time. Our experimental analysis on large-scale user data shows that our model can be used to predict future songs a user will likely listen to, both in the short and long term.

     

Recommending people to people: the nature of reciprocal recommenders with a case study in online dating

People-to-people recommenders constitute an important class of recommender systems. Examples include online dating, where people have the common goal of finding a partner, and employment websites where one group of users needs to find a job (employer) and another group needs to find an employee. People-to-people recommenders differ from the traditional items-to-people recommenders as they must satisfy both parties; we call this type of recommender reciprocal. This article is the first to present a comprehensive view of this important recommender class. We first identify the characteristics of reciprocal recommenders and compare them with traditional recommenders, which are widely used in e-commerce websites. We then present a series of studies and evaluations of a content-based reciprocal recommender in the domain of online dating. It uses a large dataset from a major online dating website. We use this case study to illustrate the distinctive requirements of reciprocal recommenders and highlight important challenges, such as the need to avoid bad recommendations since they may make users to feel rejected. Our experiments indicate that, by considering reciprocity, the rate of successful connections can be significantly improved. They also show that, despite the existence of rich explicit profiles, the use of implicit profiles provides more effective recommendations. We conclude with a discussion, linking our work in online dating to the many other domains that require reciprocal recommenders. Our key contributions are the recognition of the reciprocal recommender as an important class of recommender, the identification of its distinctive characteristics and the exploration of how these impact the recommendation process in an extensive case study in the domain of online dating.     

Exploiting the wisdom of social connections to make personalized recommendations on scholarly articles

Existing scholarly publication recommenders were designed to aid researchers, as well as ordinary users, in discovering pertinent literature in diverse academic fields. These recommenders, however, often (i) depend on the availability of users’ historical data in the form of ratings or access patterns, (ii) generate recommendations pertaining to users’ (articles included in their) profiles, as oppose to their current research interests, or (iii) fail to analyze valuable user-generated data at social sites that can enhance their performance. To address these design issues, we propose PReSA, a personalized recommender on scholarly articles. PReSA recommends articles bookmarked by the connections of a user U on a social bookmarking site that are not only similar in content to a target publication P currently of interest to U but are also popular among U’s connections. PReSA (i) relies on the content-similarity measure to identify potential academic publications to be recommended and (ii) uses only information readily available on popular social bookmarking sites to make recommendations. Empirical studies conducted using data from CiteULike have verified the efficiency and effectiveness of (the recommendation and ranking strategies of) PReSA, which outperforms a number of existing (scholarly publication) recommenders.     

Adaptive website recommendations with AWESOME

Recommendations are crucial for the success of large websites. While there are many ways to determine recommendations, the relative quality of these recommenders depends on many factors and is largely unknown. We present the architecture and implementation of AWESOME (Adaptive website recommendations), a data warehouse-based recommendation system. It allows the coordinated use of a large number of recommenders to automatically generate website recommendations. Recommendations are dynamically selected by efficient rule-based approaches utilizing continuously measured user feedback on presented recommendations. AWESOME supports a completely automatic generation and optimization of selection rules to minimize website administration overhead and quickly adapt to changing situations. We propose a classification of recommenders and use AWESOME to comparatively evaluate the relative quality of several recommenders for a sample website. Furthermore, we propose and evaluate several rule-based schemes for dynamically selecting the most promising recommendations. In particular, we investigate two-step selection approaches that first determine the most promising recommenders and then apply their recommendations for the current situation. We also evaluate one-step schemes that try to directly determine the most promising recommendations.     

Enabling topic-level trust for collaborative information sharing

As a consequence of the exponential growth of Internet and its services, including social applications fostering collaboration on the Web, information sharing had become pervasive. This caused a crescent need of more powerful tools to help users with the task of selecting interesting resources. Recommender systems have emerged as a solution to evaluate the quality of massively user-generated contents in open environments and provide recommendations based not only on the user interests but also on the opinions of people with similar tastes. In addition to interest similarity, however, trustworthiness is a factor that recommenders have to consider in the selection of reliable peers for collaboration. Most approaches in this regard estimates trust base on global user profile similarity or history of exchanged opinions. In this paper, we propose a novel approach for agent-based recommendation in which trust is independently learned and evolved for each pair of interest topics two users have in common. Experimental results show that agents learning who to trust about certain topics reach better levels of precision than considering interest similarity exclusively.     

Hybrid Recommender Systems: Survey and Experiments

Recommender systems represent user preferences for the purpose of suggesting items to purchase or examine. They have become fundamental applications in electronic commerce and information access, providing suggestions that effectively prune large information spaces so that users are directed toward those items that best meet their needs and preferences. A variety of techniques have been proposed for performing recommendation, including content-based, collaborative, knowledge-based and other techniques. To improve performance, these methods have sometimes been combined in hybrid recommenders. This paper surveys the landscape of actual and possible hybrid recommenders, and introduces a novel hybrid, EntreeC, a system that combines knowledge-based recommendation and collaborative filtering to recommend restaurants. Further, we show that semantic ratings obtained from the knowledge-based part of the system enhance the effectiveness of collaborative filtering.     

Applications of wavelet data reduction in a recommender system

Memory-based collaborative filtering (CF) recommender systems have emerged as an effective technique for information filtering. CF recommenders are being widely adopted for e-commerce applications to assist users in finding and selecting items of interest. As a result, the scalability of CF recommenders presents a significant challenge; one that is particularly resilient because the volume of data these systems utilize will continue to increase over time. This paper examines the impact of discrete wavelet transformation (DWT) as an approach to enhance the scalability of memory-based collaborative filtering recommender systems. In particular, a wavelet transformation methodology is proposed and applied to both synthetic and real-world recommender ratings. For experimental purposes, the DWT methodology’s effect on predictive accuracy and calculation speed is evaluated to compare recommendation quality and performance.