Jonghyun Han

A context management architecture for large-scale smart environments

Context-aware architecture collects various context data from heterogeneous sensors and provides an intelligent service by exploiting the collected data. In this article we explain the generalized context-aware software architecture for heterogeneous smart environments. The proposed architecture integrates large-scale contexts from multiple heterogeneous sensors, and makes a semantic decision by fusing and reasoning about the collected contexts. Moreover, we discuss a designed architecture that manages communities between large numbers of heterogeneous information entities and enhances intelligence abilities.

Adaptive landmark recommendations for travel planning

When travelers plan trips, landmark recommendation systems that consider the trip properties will conveniently aid travelers in determining the locations they will visit. Because interesting locations may vary based on the traveler and the situation, it is important to personalize the landmark recommendations by considering the traveler and the trip. In this paper, we propose an approach that adaptively recommends clusters of landmarks using geo-tagged social media. We first examine the impact of a trips spatial and temporal properties on the distribution of popular places through large-scale data analyses. In our approach, we compute the significance of landmarks for travelers based on their trips spatial and temporal properties. Next, we generate clusters of landmark recommendations, which have similar themes or are contiguous, using travel trajectory histories. Landmark recommendation performances based on our approach are evaluated against several baseline approaches. Our approach results in increased accuracy and satisfaction compared with the baseline approaches. Through a user study, we also verify that our approach is applicable to lesser-known places and reflects local events as well as seasonal changes.

Adaptive content recommendation for mobile users: Ordering recommendations using a hierarchical context model with granularity

Retrieving timely and relevant information on-site is an important task for mobile users. A context-aware system can understand a users information needs and thus select contents according to relevance. We propose a context-dependent search engine that represents user context in a knowledge-based context model, implemented in a hierarchical structure with granularity information. Search results are ordered based on semantic relevance computed as similarity between the current context and tags of search results. Compared against baseline algorithms, the proposed approach enhances precision by 22% and pooled recall by 17%. The use of size-based granularity to compute similarity makes the approach more robust against changes in the context model in comparison to graph-based methods, facilitating import of existing knowledge repositories and end-user defined vocabularies (folksonomies). The reasoning engine being light-weight, privacy protection is ensured, as all user information is processed locally on the users phone without requiring communication with an external server.

Context-based microblog browsing for mobile users

Mobile microblog browsing is inconvenient due to the limitations of mobile devices. Therefore, it is important to effectively retrieve relevant and timely microblog content that caters to the information need of mobile users. In this paper, we first present the findings from a large-scale study on the relationship between microblog content and user context. Then, we show a system that detects local microblog topics, estimates user interests and selects user-preferred topics. The system employs user context to detect microblog topics and post-processes the topics for finding user-preferred content. We exploit time, location, browsing history, social relationship and activity as user context. The effectiveness of our approach is evaluated against several baseline algorithms for investigating the impact of user context on the relevance of retrieved topics. According to our experimental results, the approach enhances the relevance of topics by 24%, compared to the baseline approaches. Thus, we expect that the proposed approach is helpful in advancing mobile information retrieval.

Characterizing user interest using heterogeneous media

It is often hard to accurately estimate interests of social media users because their messages do not have additional information, such as a category. In this paper, we propose an approach that estimates user interest from social media to provide personalized services. Our approach employs heterogeneous media to map social media onto categories. To describe the categories, we propose a hybrid method that integrates a topic model with TF-ICF for extracting both explicitly presented and implicitly latent features. Our evaluation result shows that it gives the highest performance, compared to other approaches. Thus, we expect that the proposed approach is helpful in advancing personalization of social media services.

Visualization and management of u-contents for ubiquitous VR

Ubiquitous Virtual Reality, where ubiquitous computing meets mixed reality, is coming to our lives based on recent developments in the two fields. In this paper, we focus on the conceptual properties of contents including definition rather than infrastructures or algorithms for Ubiquitous Virtual Reality. For this purpose, we define u-Content and its descriptor with three conceptual key properties: u-Realism, u-Intelligence, and u-Mobility. Then we address the overall scheme of the descriptor with a Context-aware Augmented Reality Toolkit for visualization and management. We also show how the proposed concept is applied in the recent applications.

In silico re-identification of properties of drug target proteins

BackgroundComputational approaches in the identification of drug targets are expected to reduce time and effort in drug development. Advances in genomics and proteomics provide the opportunity to uncover properties of druggable genomes. Although several studies have been conducted for distinguishing drug targets from non-drug targets, they mainly focus on the sequences and functional roles of proteins. Many other properties of proteins have not been fully investigated.MethodsUsing the DrugBank (version 3.0) database containing nearly 6,816 drug entries including 760 FDA-approved drugs and 1822 of their targets and human UniProt/Swiss-Prot databases, we defined 1578 non-redundant drug target and 17,575 non-drug target proteins. To select these non-redundant protein datasets, we built four datasets (A, B, C, and D) by considering clustering of paralogous proteins.ResultsWe first reassessed the widely used properties of drug target proteins. We confirmed and extended that drug target proteins (1) are likely to have more hydrophobic, less polar, less PEST sequences, and more signal peptide sequences higher and (2) are more involved in enzyme catalysis, oxidation and reduction in cellular respiration, and operational genes. In this study, we proposed new properties (essentiality, expression pattern, PTMs, and solvent accessibility) for effectively identifying drug target proteins. We found that (1) drug targetability and protein essentiality are decoupled, (2) druggability of proteins has high expression level and tissue specificity, and (3) functional post-translational modification residues are enriched in drug target proteins. In addition, to predict the drug targetability of proteins, we exploited two machine learning methods (Support Vector Machine and Random Forest). When we predicted drug targets by combining previously known protein properties and proposed new properties, an F-score of 0.8307 was obtained.ConclusionsWhen the newly proposed properties are integrated, the prediction performance is improved and these properties are related to drug targets. We believe that our study will provide a new aspect in inferring drug-target interactions.

Analyzing social media friendship for personalization

Since social media users have various purposes such as tightening friendship and obtaining information, it might be easier to model a users interest and to provide personalized information if the users purpose can be inferred. In this paper, we analyze the friendship of Twitter users and its effects on Twitter usage. According to our analysis, although the number of offline friends is smaller than that of online friends, a user more actively responds to the microblogs posted by the offline friends. We expect that our analysis is helpful to model a users social behavior and interest.