Search results

This paper aims to identify, evaluate and integrate the findings of all relevant and high-quality individual studies addressing one or more research questions about recommender systems and performing a comprehensive study of empirical research on recommender systems that have been divided into five main categories. To achieve this aim, the authors use systematic literature review (SLR) as a powerful method to collect and critically analyze the research papers. Also, the authors discuss the selected recommender systems and its main techniques, as well as their benefits and drawbacks in general.

In this paper, the SLR method is utilized with the aim of identifying, evaluating and integrating the findings of all relevant and high-quality individual studies addressing one or more research questions about recommender systems and performing a comprehensive study of empirical research on recommender systems that have been divided into five main categories. Also, the authors discussed recommender system and its techniques in general without a specific domain.

The major developments in categories of recommender systems are reviewed, and new challenges are outlined. Furthermore, insights on the identification of open issues and guidelines for future research are provided. Also, this paper presents the systematical analysis of the recommender system literature from 2005. The authors identified 536 papers, which were reduced to 51 primary studies through the paper selection process.

This survey will directly support academics and practical professionals in their understanding of developments in recommender systems and its techniques.

Recommender agents are used to make recommendations of interesting items in a wide variety of application domains, such as web page recommendation, music, e‐commerce, movie recommendation, tourism, restaurant recommendation, among others. Despite the various and different domains in which recommender agents are used and the variety of approaches they use to represent user interests and make recommendations, there is some functionality that is common to all of them, such as user model management and recommendation of interesting items. This paper aims at generalizing these common behaviors into a framework that enables developers to reuse recommender agents' main characteristics in their own developments.

This work presents a framework for recommendation that provides the control structures, the data structures and a set of algorithms and metrics for different recommendation methods. The proposed framework acts as the base design for recommender agents or applications that want to add the already modeled and implemented capabilities to their own functionality. In contrast with other proposals, this framework is designed to enable the integration of diverse user models, such as demographic, content‐based and item‐based. In addition to the different implementations provided for these components, new algorithms and user model representations can be easily added to the proposed approach. Thus, personal agents originally designed to assist a single user can reuse the behavior implemented in the framework to expand their recommendation strategies.

The paper describes three different recommender agents built by materializing the proposed framework: a movie recommender agent, a tourism recommender agent, and a web page recommender agent. Each agent uses a different recommendation approach. PersonalSearcher, an agent originally designed to suggest interesting web pages to a user, was extended to collaboratively assist a group of users using content‐based algorithms. MovieRecommender recommends interesting movies using an item‐based approach and Traveller suggests holiday packages using demographic user models. Findings encountered during the development of these agents and their empirical evaluation are described here.

The advantages of the proposed framework are twofold. On the one hand, the functionality provided by the framework enables the development of recommender agents without the need for implementing its whole set of capabilities from scratch. The main processes and data structures of recommender agents are already implemented. On the other hand, already existing agents can be enhanced by incorporating the functionality provided by the recommendation framework in order to act collaboratively.

The information needs of the users of literature database systems often come from the task at hand, which is short term and can be represented as a small number of articles. Previous works on recommending articles to satisfy users’ short-term interests have utilized article content, usage logs, and more recently, coauthorship networks. The usefulness of coauthorship has been demonstrated by some research works, which, however, tend to adopt a simple coauthorship network that records only the strength of coauthorships. The purpose of this paper is to enhance the effectiveness of coauthorship-based recommendation by incorporating scholars’ collaboration topics into the coauthorship network.

The authors propose a latent Dirichlet allocation (LDA)-coauthorship-network-based method that integrates topic information into the links of the coauthorship networks using LDA, and a task-focused technique is developed for recommending literature articles.

The experimental results using information systems journal articles show that the proposed method is more effective than the previous coauthorship network-based method over all scenarios examined. The authors further develop a hybrid method that combines the results of content-based and LDA-coauthorship-network-based recommendations. The resulting hybrid method achieves greater or comparable recommendation effectiveness under all scenarios when compared to the content-based method.

This paper makes two contributions. The authors first show that topic model is indeed useful and can be incorporated into the construction of coaurthoship-network to improve literature recommendation. The authors subsequently demonstrate that coauthorship-network-based and content-based recommendations are complementary in their hit article rank distributions, and then devise a hybrid recommendation method to further improve the effectiveness of literature recommendation.

With the rapid increase in internet use, most people tend to purchase books through online stores. Several such stores also provide book recommendations for buyer convenience, and both collaborative and content-based filtering approaches have been widely used for building these recommendation systems. However, both approaches have significant limitations, including cold start and data sparsity. To overcome these limitations, this study aims to investigate whether user satisfaction can be predicted based on easily accessible book descriptions.

The authors collected a large-scale Kindle Books data set containing book descriptions and ratings, and calculated whether a specific book will receive a high rating. For this purpose, several feature representation methods (bag-of-words, term frequency–inverse document frequency [TF-IDF] and Word2vec) and machine learning classifiers (logistic regression, random forest, naive Bayes and support vector machine) were used.

The used classifiers show substantial accuracy in predicting reader satisfaction. Among them, the random forest classifier combined with the TF-IDF feature representation method exhibited the highest accuracy at 96.09%.

This study revealed that user satisfaction can be predicted based on book descriptions and shed light on the limitations of existing recommendation systems. Further, both practical and theoretical implications have been discussed.

To retain consumer attention and increase purchasing rates, many e-commerce vendors have adopted content-based recommender systems. However, apart from text-based documents, there is little theoretical background guiding element selection, resulting in a limited content analysis problem. Another inherent problem is overspecialization. The purpose of this paper is to establish a value-based recommendation methodology for identifying favorable attributes, benefits, and values on the basis of means-end chain theory. The identified elements and the relationships between them were utilized to construct a recommender system without incurring either problem.

This study adopted soft laddering and content analysis to collect popular elements. The relationships between the elements were established by using a hard laddering online questionnaire. The elements and the relationships were utilized to build a hierarchical value map (HVM). A mathematical model was then devised on the basis of the HVM to predict user preferences of attributes.

The results of a performance comparison showed that the proposed method outperformed the content-based attribute recommendation method and a hybrid method by 39 and 68 percent, respectively.

Although hybrid methods have been proposed to resolve the problem of overspecialization in content-based recommender systems, such methods have incurred “cold start” and “sparsity” problems. The proposed method can provide recommendations without causing these problems while outperforming the content-based and hybrid approaches.

In a large‐scale digital library, it is essential to recommend a small number of useful and related articles to users. In this paper, a literature recommendation framework for digital libraries is proposed that dynamically provides recommendations to an active user when browsing a new article. This framework extends our previous work that considers only Web usage data by utilizing content information of articles when making recommendations. Methods that make use of pure content data, pure Web usage data, and both content and usage data are developed and compared using the data collected from our university's electronic thesis and dissertation (ETD) system. The experimental results demonstrate that content data and usage data are complements of each other and hybrid methods that take into account of both types of information tend to achieve more accurate recommendations.

The purpose of this paper is to describe a new ontological content‐based filtering method for ranking the relevance of items for readers of news items, and its evaluation. The method has been implemented in ePaper, a personalised electronic newspaper prototype system. The method utilises a hierarchical ontology of news; it considers common and related concepts appearing in a user's profile on the one hand, and in a news item's profile on the other hand, and measures the “hierarchical distances” between these concepts. On that basis it computes the similarity between item and user profiles and rank‐orders the news items according to their relevance to each user.

The paper evaluates the performance of the filtering method in an experimental setting. Each participant read news items obtained from an electronic newspaper and rated their relevance. Independently, the filtering method is applied to the same items and generated, for each participant, a list of news items ranked according to relevance.

The results of the evaluations revealed that the filtering algorithm, which takes into consideration hierarchically related concepts, yielded significantly better results than a filtering method that takes only common concepts into consideration. The paper determined a best set of values (weights) of the hierarchical similarity parameters. It also found out that the quality of filtering improves as the number of items used for implicit updates of the profile increases, and that even with implicitly updated profiles, it is better to start with user‐defined profiles.

The proposed content‐based filtering method can be used for filtering not only news items but items from any domain, and not only with a three‐level hierarchical ontology but any‐level ontology, in any language.

The purpose of this paper is to develop a cross‐language personalized recommendation model based on web log mining, which can recommend academic articles, in different languages, to users according to their demands.

The proposed model takes advantage of web log data archived in digital libraries and learns user profiles by means of integration analysis of a user's multiple online behaviors. Moreover, keyword translation was carried out to eliminate language dissimilarity between user and item profiles. Finally, article recommendation can be achieved using various existing algorithms.

The proposed model can recommend articles in different languages to users according to their demands, and the integration analysis of multiple online behaviors can help to better understand a user's interests.

This study has a significant implication for digital libraries in non‐English countries, since English is the most popular language in current academic articles and it is a very common phenomenon for users in these countries to obtain literatures presented by more than one language. Furthermore, this approach is also useful for other text‐based item recommendation systems.

A lot of research work has been done in the personalized recommendation area, but few works have discussed the recommendation problem under multiple linguistic circumstances. This paper deals with cross‐language recommendation and, moreover, the proposed model puts forward an integration analysis method based on multiple online behaviors to understand users' interests, which can provide references for other recommendation systems in the digital age.

Recommender system approaches such as collaborative and content-based filtering rely on user ratings and product descriptions to recommend products. More recently, recommender system research has focussed on exploiting knowledge from user-generated content such as product reviews to enhance recommendation performance. The purpose of this paper is to show that the performance of a recommender system can be enhanced by integrating explicit knowledge extracted from product reviews with implicit knowledge extracted from analysis of consumer’s purchase behaviour.

The authors introduce a sentiment and preference-guided strategy for product recommendation by integrating not only explicit, user-generated and sentiment-rich content but also implicit knowledge gleaned from users’ product purchase preferences. Integration of both of these knowledge sources helps to model sentiment over a set of product aspects. The authors show how established dimensionality reduction and feature weighting approaches from text classification can be adopted to weight and select an optimal subset of aspects for recommendation tasks. The authors compare the proposed approach against several baseline methods as well as the state-of-the-art better method, which recommends products that are superior to a query product.

Evaluation results from seven different product categories show that aspect weighting and selection significantly improves state-of-the-art recommendation approaches.

The proposed approach recommends products by analysing user sentiment on product aspects. Therefore, the proposed approach can be used to develop recommender systems that can explain to users why a product is recommended. This is achieved by presenting an analysis of sentiment distribution over individual aspects that describe a given product.

This paper describes a novel approach to integrate consumer purchase behaviour analysis and aspect-level sentiment analysis to enhance recommendation. In particular, the authors introduce the idea of aspect weighting and selection to help users identify better products. Furthermore, the authors demonstrate the practical benefits of this approach on a variety of product categories and compare the approach with the current state-of-the-art approaches.

Exploratory learning is regarded as an important ability for developing knowledge from open environments. During the exploration, learners not only need to acquire new information based on their current interests, but also they need to form new perspectives by incorporating new knowledge into their previous knowledge. This paper seeks to address these issues.

To this end, this paper proposes an approach that includes a concept association bank to recommend related concepts in a domain based on the goal of an exploration. By doing so, learners' knowledge can be expanded beyond their current understanding. An experiment was conducted to investigate how the proposed approach facilitated the learners' exploration.

The results indicated that the concept association bank is a useful mechanism to help learners gain new understanding, including providing exploration directions, reducing complexity and cognitive load, facilitating data‐ and goal‐driven exploration strategies, and commenting on new understanding. The implications of these results are discussed.

Current recommendation systems emphasise a data‐driven strategy, which seeks isolated pieces of information, instead of suggesting directions related to their exploration goal. The problem with such an approach is that learners' exploration will be limited by their existing knowledge. Thus, this paper presents an approach to support both data‐ and goal‐driven strategies.