Kyung-Yong Jung

Hybrid Collaborative Filtering and Content-Based Filtering for Improved Recommender System

The growth of the Internet has resulted in an increasing need for personalized information systems. The paper describes an autonomous agent, WebBot: Web Robot Agent, which integrates with the web and acts as a personal recommender system that cooperates with the user on identifying interesting pages. Hybrid components from collaborative filtering and content-based filtering, a hybrid recommender system can overcome traditional shortcomings. In this paper, we present an effective hybrid collaborative filtering and content-based filtering for improved recommender system. Experimental results indicate the hybrid collaborative filtering and content-based filtering better than collaborative, content-based, and combined filtering approach.

Prediction of User Preference in Recommendation System Using Associative User Clustering and Bayesian Estimated Value

The user predicting preference method using a collaborative filtering (CF) does not only reflect any contents about items but also solve the sparsity and first-rater problem. In this paper, we suggest the method of prediction by using associative user clustering and Bayesian estimated value to complement the problems of the current collaborative filtering system. The Representative Attribute-Neighborhood is for an active user to select the nearest neighbors who have similar preference through extracting the representative attributes that most affects the preference. Associative user behavior pattern 3_UB(associative users are composed of 3-users) is clustered according to the genre through Association Rule Hypergraph Partitioning Algorithm, and new users are classified into one of these genres by Naive Bayes classifier. Besides, to get the similarity between users belonged to the classified genre and new users, this paper allows the different estimated values to items which users evaluated through Naive Bayes learning. We evaluate our method on a large CF database of user rating and it significantly outperforms the previous proposed method.

Interleukin-18 as an efficient marker for remission and follow-up in patients with inactive adult-onset Still’s disease

Objectives: Diagnosis of adult-onset Still’s disease (AOSD) is difficult because of a lack of pathognomonic findings and markers. The aim of this study was to investigate the efficacy of interleukin (IL)-18 and free IL-18 in the diagnosis and follow-up of patients with AOSD. Method: Levels of inflammatory cytokines, IL-18, IL-18 binding protein (IL-18BP), and free IL-18 were compared in 80 AOSD patients and 90 controls. The AOSD patients were divided into active and inactive groups according to disease activity, and the inactive patients were subdivided into a remission subgroup and a low disease activity subgroup. We compared erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), ferritin, IL-18, and free IL-18 as disease activity markers in the AOSD patients. Serial serum levels of activity markers were measured in 52 of the 80 AOSD patients at 3- to 6-month intervals. Results:There were significantly higher levels of IL-18 and free IL-18 in the AOSD patients than in the controls. IL-18 and free IL-18 were significantly higher in the active group than the inactive group (p < 0.001 for all). Unlike other activity markers, IL-18 and free IL-18 levels in the low disease activity subgroup were significantly higher than those in the remission subgroup within the inactive group (p = 0.004 and 0.005, respectively). During serial follow-up, ferritin and IL-18 showed a significant decrease in the responder and remission subgroup. Conclusions: IL-18 might be an efficient marker for diagnosis and follow-up of AOSD and might also be a useful predictor of remission, especially in clinically inactive patients.

Discovery of User Preference in Personalized Design Recommender System through Combining Collaborative Filtering and Content Based Filtering

More and more recommender systems build close relationships with their users by adapting to their needs and therefore providing a personal experience. One aspect of personalization is the recommendation and presentation of information and products so that users can access the recommender system more efficiently. However, powerful filtering technology is required in order to identify relevant items for each user. In this paper we describe how collaborative filtering and content-based filtering can be combined to provide better performance for information filtering. We propose the personalized design recommender system of textile design applying both technologies as one of the methods in the material development centered on customer’s sensibility and preference. Finally, we plan to conduct empirical applications to verify the adequacy and the validity of our personalized design recommender system.

FDRAS: Fashion Design Recommender Agent System Using the Extraction of Representative Sensibility and the Two-Way Combined Filtering on Textile

It is important for the strategy of product sales to investigate the customer’s sensibility and preference degree in the environment that the process of material development has been changed focusing on the customer center. In this paper we identify collaborative filtering and content-based filtering as independent technologies for information filtering. We propose the Fashion Design Recommender Agent System of textile design applying two-way combined filtering technologies as one of methods in the material development centered on customer’s representative sensibility and preference. We build the database founded on the sensibility adjective to develop textile design by extracting the representative sensibility adjective form user’s sensibility and preference about textiles. Our system recommends textile designs to a customer who has a similar propensity about textile. Ultimately, this paper suggests empirical applications to verify the adequacy and the validity on this system.

Creating User-Adapted Design Recommender System through Collaborative Filtering and Content Based Filtering

Information filtering is an important technology for the creation of recommender system, which are adapted to the user’s needs. In this paper we identify collaborative filtering and content-based filtering as independent technologies for information filtering. We propose the user-adapted design recommender system of textile design applying both technologies as one of methods in the material development centered on customer’s sensibility and preference. Ultimately, this paper suggests empirical applications to verify the adequacy and the validity on this system with the development of design recommender system.

Development of Design Recommender System Using Collaborative Filtering

It is an important strategy to investigate customer’s sensibility and preference in the merchandise environment changing to the user oriented. We propose the design recommender system, which exposes its collection in a personalized way by the use of collaborative filtering and representative sensibility adjective on textile design. We developed the multi-users interface tool that can suggest designs according to the user’s needs in the design industry. In this paper, we adapt collaborative filtering to recommend design to a user who has a similar propensity about designs. And we validate our design recommender system according to three algorithms in off-line experiments. Design merchandizing may meet the consumer’s needs more exactly and easily with this system.

Dynamic Single Path Routing Mechanism for Reliability and Energy-Efficiency in a Multi Hop Sensor Network

What are important in wireless sensor networks are reliable data transmission, energy efficiency of each node, and the maximization of network life through the distribution of load among the nodes. The present study proposed DSPR, a dynamic unique path routing machanism that considered these requirements in wireless sensor networks. In DSPR, data is transmitted through a dynamic unique path, which has the least cost calculated with the number of hops from each node to the sink, and the average remaining energy. At that time, each node monitors its transmission process and if a node detects route damage it changes the route dynamically, referring to the cost table, and by doing so, it enhances the reliability of the network and distributes energy consumption evenly among the nodes. In addition, when the network topology is changed, only the part related to the change is restructured dynamically instead of restructuring the entire network, and the life of the network is extended by inhibiting unnecessary energy consumption in each node as much as possible. In the results of our experiment, the proposed DSPR increased network life by minimizing energy consumption of the nodes and improved the reliability and energy efficiency of the network.

Design of Ubiquitous Music Recommendation System Using MHMM

The existing music search and recommendation systems obtain results through query or answer and recommend music using data mining techniques. However, it is not possible to provide active services that satisfy customers in smart home environments because these systems consider only static information in Web environments. In order to solve these problems, this paper attempts to define context information to use select music and design a ubiquitous music recommendation system that is suited to a users interests and preferences using hidden Markov model for music items. The recommendation system used in this study uses an OSGi framework to recognize context information and increase satisfaction of service.

Automatic Preference Mining through Learning User Profile with Extracted Information

Previous Bayesian classification has a problem because of reflecting semantic relation accurately in expressing characteristic of web pages. To resolve this problem, this paper proposes automatic preference mining through learning user profile with extracted information. Apriori algorithm extracts characteristic of web pages in form of association words that reflects semantic relation and it mines association words from learning the ontological user profile. Our prototype personalized movie recommender system, WebBot, extracts information about movies from web pages to recommend titles based on training movie set supplied by an individual user. The proposed method was tested in database that users estimated the preference about web pages, and certified that was more efficient than existent methods.