• Smart Media Journal
• /
• v.12 no.4
• /
• pp.27-40
• /
• 2023

Many researchers have investigated ways to enhance recommender systems by integrating heterogeneous data to address the data sparsity problem. However, only a few studies have successfully integrated heterogeneous data using knowledge graph. Additionally, most of the knowledge graphs built in these studies only incorporate explicit relationships between entities and lack additional information. Therefore, we propose a method for expanding knowledge graphs by using deep learning to model latent relationships between heterogeneous data from multiple knowledge bases. Our extended knowledge graph enhances the quality of entity features and ultimately increases the accuracy of predicted user preferences. Experiments using real music data demonstrate that the expanded knowledge graph leads to an increase in recommendation accuracy when compared to the original knowledge graph.

• Journal of Intelligence and Information Systems
• /
• v.21 no.1
• /
• pp.119-142
• /
• 2015

With the explosive growth in the volume of information, Internet users are experiencing considerable difficulties in obtaining necessary information online. Against this backdrop, ever-greater importance is being placed on a recommender system that provides information catered to user preferences and tastes in an attempt to address issues associated with information overload. To this end, a number of techniques have been proposed, including content-based filtering (CBF), demographic filtering (DF) and collaborative filtering (CF). Among them, CBF and DF require external information and thus cannot be applied to a variety of domains. CF, on the other hand, is widely used since it is relatively free from the domain constraint. The CF technique is broadly classified into memory-based CF, model-based CF and hybrid CF. Model-based CF addresses the drawbacks of CF by considering the Bayesian model, clustering model or dependency network model. This filtering technique not only improves the sparsity and scalability issues but also boosts predictive performance. However, it involves expensive model-building and results in a tradeoff between performance and scalability. Such tradeoff is attributed to reduced coverage, which is a type of sparsity issues. In addition, expensive model-building may lead to performance instability since changes in the domain environment cannot be immediately incorporated into the model due to high costs involved. Cumulative changes in the domain environment that have failed to be reflected eventually undermine system performance. This study incorporates the Markov model of transition probabilities and the concept of fuzzy clustering with CBCF to propose predictive clustering-based CF (PCCF) that solves the issues of reduced coverage and of unstable performance. The method improves performance instability by tracking the changes in user preferences and bridging the gap between the static model and dynamic users. Furthermore, the issue of reduced coverage also improves by expanding the coverage based on transition probabilities and clustering probabilities. The proposed method consists of four processes. First, user preferences are normalized in preference clustering. Second, changes in user preferences are detected from review score entries during preference transition detection. Third, user propensities are normalized using patterns of changes (propensities) in user preferences in propensity clustering. Lastly, the preference prediction model is developed to predict user preferences for items during preference prediction. The proposed method has been validated by testing the robustness of performance instability and scalability-performance tradeoff. The initial test compared and analyzed the performance of individual recommender systems each enabled by IBCF, CBCF, ICFEC and PCCF under an environment where data sparsity had been minimized. The following test adjusted the optimal number of clusters in CBCF, ICFEC and PCCF for a comparative analysis of subsequent changes in the system performance. The test results revealed that the suggested method produced insignificant improvement in performance in comparison with the existing techniques. In addition, it failed to achieve significant improvement in the standard deviation that indicates the degree of data fluctuation. Notwithstanding, it resulted in marked improvement over the existing techniques in terms of range that indicates the level of performance fluctuation. The level of performance fluctuation before and after the model generation improved by 51.31% in the initial test. Then in the following test, there has been 36.05% improvement in the level of performance fluctuation driven by the changes in the number of clusters. This signifies that the proposed method, despite the slight performance improvement, clearly offers better performance stability compared to the existing techniques. Further research on this study will be directed toward enhancing the recommendation performance that failed to demonstrate significant improvement over the existing techniques. The future research will consider the introduction of a high-dimensional parameter-free clustering algorithm or deep learning-based model in order to improve performance in recommendations.

• KIISE Transactions on Computing Practices
• /
• v.24 no.2
• /
• pp.93-98
• /
• 2018

Recommendation systems are used to provide items of interests for users to maximize a company's profit. Matrix factorization is frequently used by recommendation systems, based on an incomplete user-item rating matrix. However, as the number of items and users increase, it becomes difficult to make accurate recommendations due to the sparsity of data. To overcome this drawback, the use of text data related to items was recently suggested for matrix factorization algorithms. Furthermore, a word-level convolutional neural network was shown to be effective in the process of extracting the word-level features from the text data among these kinds of matrix factorization algorithms. However, it involves a large number of parameters to learn in the word-level convolutional neural network. Thus, we propose a matrix factorization algorithm which utilizes a character-level convolutional neural network with which to extract the character-level features from the text data. We also conducted a performance study with real-life datasets to show the effectiveness of the proposed matrix factorization algorithm.

• The Journal of Society for e-Business Studies
• /
• v.14 no.4
• /
• pp.47-57
• /
• 2009

Collaborative tagging systems allow users to attach tags to diverse sharable contents in social networks. These tags provide usefulness in reusing the contents for all community members as well as their creators. Three-dimensional data composed of users, items, and tags are used in the collaborative tag-based recommendation. They are generally more voluminous and sparse than two-dimensional data composed of users and items. Therefore, there are many difficulties in applying existing collaborative filtering methods directly to them. Latent factor models, which are also successful in the area of collaborative filtering recently, discover latent features(factors) for explaining observed values and solve problems based on the features. However, establishing the models require much time and efforts. In order to apply the latent factor models to three-dimensional collaborative filtering data, we have to overcome the difficulty of establishing them. This paper proposes various methods for determining preferences of users to items via establishing an intuitive model by assuming tags used for items as latent factors to users and items respectively. They are compared using real data for concluding desirable directions.

• Journal of Intelligence and Information Systems
• /
• v.13 no.3
• /
• pp.119-140
• /
• 2007

Recommender systems enable a user to decide which information is interesting and valuable in our world of information overload. As the recent studies of distributed computing environment have been progressing actively, recommender systems, most of which were centralized, have changed toward a peer-to-peer approach. Collaborative Filtering (CF), one of the most successful technologies in recommender systems, presents several limitations, namely sparsity, scalability, cold start, and the shilling problem, in spite of its popularity. The move from centralized systems to distributed approaches can partially improve the issues; distrust of recommendation and abuses of personal information. However, distributed systems can be vulnerable to attackers, who may inject biased profiles to force systems to adapt their objectives. In this paper, we consider both effective CF in P2P environment in order to improve overall performance of system and efficient solution of the problems related to abuses of personal data and attacks of malicious users. To deal with these issues, we propose a multi-agent framework for a distributed CF focusing on the trust relationships between individuals, i.e. web of trust. We employ an agent-based approach to improve the efficiency of distributed computing and propagate trust information among users with effect. The experimental evaluation shows that the proposed method brings significant improvement in terms of the distributed computing of similarity model building and the robustness of system against malicious attacks. Finally, we are planning to study trust propagation mechanisms by taking trust decay problem into consideration.

• The Journal of Korean Association of Computer Education
• /
• v.9 no.1
• /
• pp.71-80
• /
• 2006

As e-commerce developing rapidly, it is becoming a research focus about how to find customer's behavior patterns and realize commerce intelligence by use of Web mining technology. One of the most successful and widely used technologies for building personalization and goods recommendation system is collaborating filtering. However, collaborative filtering have serious data sparsity problem. Traditional association rule does not consider user's interests or preferences to provide a user with specific personalized service.In this paper, we propose an goods recommendation system, which is integrated an collaborative filtering algorithm with item-to-item corelation and an improved Apriori algorithm. This system has user's interests or preferences ro provide a user with specific personalized service.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.12
• /
• pp.5103-5115
• /
• 2015

In radar imaging, a target is usually consisted of a few strong scatterers which are sparsely distributed. In this paper, an improved sparse signal recovery algorithm based on smoothed l₀ (SL0) norm method is proposed to achieve high resolution ISAR imaging with limited pulse numbers. Firstly, one new smoothed function is proposed to approximate the l₀ norm to measure the sparsity. Then a single loop step is used instead of two loop layers in SL0 method which increases the searching density of variable parameter to ensure the recovery accuracy without increasing computation amount, the cost function is undated in every loop for the next loop until the termination is satisfied. Finally, the new set of solution is projected into the feasible set. Simulation results show that the proposed algorithm is superior to the several popular methods both in terms of the reconstruction performance and computation time. Real data ISAR imaging obtained by the proposed algorithm is competitive to several other methods.

• Proceedings of the KIEE Conference
• /
• 1995.11a
• /
• pp.523-526
• /
• 1995

This paper presents an improved algorithm of optimal measurement system design with a reliability evaluation method for a large power system. The proposed algorithm is developed to cosider the observability and to achieve highest accuracy of the state estimator as well with the limited investment cost. When the effect on these dummy bus measurements is considered in the proposed algorithm the other errors in the power system is also detected and then analyzed until to achieve the limited values. By taking advantage of the matrix sparsity and the optimal bus ordering the memory and the time are successfully reduced in the P/C's and workstation's model. The improved program is successfully tested for IEEK sample system and KEPCO system with PSS/E lineflow calculated data package.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.5 no.3
• /
• pp.560-574
• /
• 2011

In this paper, we propose a hybrid method of Machine Learning (ML) algorithm and a rule-based algorithm to implement a lightweight Named Entity (NE) extraction system for Korean SMS text. NE extraction from Korean SMS text is a challenging theme due to the resource limitation on a mobile phone, corruptions in input text, need for extension to include personal information stored in a mobile phone, and sparsity of training data. The proposed hybrid method retaining the advantages of statistical ML and rule-based algorithms provides fully-automated procedures for the combination of ML approaches and their correction rules using a threshold-based soft decision function. The proposed method is applied to Korean SMS texts to extract person's names as well as location names which are key information in personal appointment management system. Our proposed system achieved 80.53% in F-measure in this domain, superior to those of the conventional ML approaches.

• Journal of Biomedical Engineering Research
• /
• v.35 no.3
• /
• pp.68-74
• /
• 2014

Recent research on mild cognitive impairment (MCI) has shown that cognitive and memory decline in this disease is accompanied by disruptive changes in the brain functional network. However, there have been no graph-theoretical studies using $^{11}C$-PIB PET data of the Alzheimer's Disease or mild cognitive impairment. In this study, we acquired $^{18}F$-FDG PET and $^{11}C$-PIB PET images of twenty-four normal aging control participants and thirty individuals with MCI from ADNI (Alzheimer's Disease Neuroimaging Initiative) database. Brain networks were constructed by thresholding binary correlation matrices using graph theoretical approaches. Both normal control and MCI group showed small-world property in $^{11}C$-PIB PET images as well as $^{18}F$-FDG PET images. $^{11}C$-PIB PET images showed significant difference between NC (normal control) and MCI over large range of sparsity values. This result will enable us to further analyze the brain using established graph-theoretical approaches for $^{11}C$-PIB PET images.