Showing papers by "Geun-Sik Jo published in 2006"

Location-based service with context data for a restaurant recommendation

Abstract: Utilizing Global Positioning System (GPS) technology, it is possible to find and recommend restaurants for users operating mobile devices. For recommending restaurants, Personal Digital Assistants or cellular phones only consider the location of restaurants. However, a user's background and environment information is assumed to be directly related to recommendation quality. In this paper, therefore, a recommender system using context information and a decision tree model for efficient recommendation is presented. This system considers location context, personal context, environment context, and user preference. Restaurant lists are obtained from location context, personal context, and environment context using the decision tree model. In addition, a weight value is used for reflecting user preferences. Finally, the system recommends appropriate restaurants to the mobile user. For this experiment, performance was verified using measurements such as k-fold cross-validation and Mean Absolute Error. As a result, the proposed system obtained an improvement in recommendation performance.

Enhanced prediction algorithm for item-based collaborative filtering recommendation

Abstract: As the Internet infrastructure has been developed, a substantial number of diverse effective applications have attempted to achieve the full potential offered by the infrastructure. Collaborative Filtering recommender system, one of the most representative systems for personalized recommendations in E-commerce on the Web, is a system assisting users in easily finding the useful information. But traditional collaborative filtering suffers some weaknesses with quality evaluation: the sparsity of the data, scalability, unreliable users. To address these issues, we have presented a novel approach to provide the enhanced prediction quality supporting the protection against the influence of malicious ratings, or unreliable users. In addition, an item-based approach is employed to overcome the sparsity and scalability problems. The proposed method combines the item confidence and item similarity, collectively called item trust using this value for online predictions. The experimental evaluation on MovieLens datasets shows that the proposed method brings significant advantages both in terms of improving the prediction quality and in dealing with malicious datasets.

Optimization of Base Stations Positioning in Mobile Networks

Abstract: Cell planning is a very complex task because many aspects should be taken into account including the traffic distribution, geography, antenna height of base stations, frequency, and so forth. This becomes more complicated when several constraints are considered such as the quality of service, maximum output power of transmitters, types of power control, and other network requirements. The problems are assignment of mobile stations to the appropriate base station and finding good locations of the base stations. This paper presents a constraint satisfaction problem model for base station location in 3G W-CDMA uplink environments. This is similar to a warehouse location problem, so it is NP-hard and consequently cannot be practically solved by exact methods for real size networks. Thus we apply constraint satisfaction techniques such as variable ordering and value ordering to get good approximate solutions. The algorithm has been implemented by using both integer programming and constraint programming.

User action based adaptive learning with weighted bayesian classification for filtering spam mail

Abstract: Nowadays, e-mail is considered one of the most important communication methods, but most users suffer from Spam mail. To solve this problem, there has been much research. The previous research showed comparatively high performance, but for adaptation of real world, it requires several improvements. First, it needs personalized learning for better performance. We cannot make a strict definition of Spam, because the definition of any context depends on each user. Second, the concept drift or interest drift problem, that is, users' interest or any context's concept, may change over time. Therefore, many Spam filtering systems are using continuous learning schemes such as adaptive learning or incremental learning. However, these systems require user feedback or rating results manually, and this inconvenience causes slow learning and performance enhancement. In this research, we developed an adaptive learning system based on an automatic weighting environment. For the automatic weight, we categorized 6 user patterns (actions) on the mailing system whose weights are automatically adapted to the learning phase. From the experiment, we will demonstrate the Bayesian classification with an adaptive learning environment. By using suggesting ideas, we will analyze the comparison result with adaptive learning. Finally, from the experiment using real world data sets, we will prove its possibility for tracking the concept and interest drift problems.

Collaborative user tracking for community organization on blogosphere: a case study of eLearning@BlogGrid

Abstract: The aim of this study is to recommend relevant information to users by organizing user communities on electronic learning environment. In this paper, we propose a weblog-based approach to modeling users during collaborative learning process. Thereby, we formulate user behaviors on blogosphere, e.g., posting articles, linking to neighbors, and interactions between neighbors. These user models are capable of being compared with others to quantify similarities between users. We apply co-occurrence analysis methods. In this study, we deploy BlogGrid platform to support information pushing service to students. Through our experimental results, we found out that average weighting measurement scheme with co-occurrence patterns from responding (e.g., comments and trackback) activities is the most significant patterns for information pushing on collaborative learning.

An agent-based multi-issue negotiation model in e-commerce

Abstract: Our paper proposes an agent-based automated negotiation model. The agents can perform an integrative negotiation with multi-issue in a one-to-many way. The negotiation protocol follows the offer-counteroffer principal, and an adapted mechanism of offer generation strategy. With the utility theory, agent could evaluate the offers and determine the following actions. In order to yield a top-quality deal and shorten the negotiation period, agents propose multiple offers, which consist of a particular combination of issue values and have the identical utility with the given utility. The experiment shows that the model ensures the participants could reach a mutually beneficial agreement in a short time.

Solving a constraint satisfaction problem for frequency assignment in low power FM broadcasting

Abstract: We present an algorithm to solve the frequency assignment problem for low power FM broadcasting. To get a good suboptimal solution, some heuristics such as k-coloring variable ordering and mostly-used value ordering rule are provided. They enforce the backtracking process in a constraint satisfaction problem, so both the search space and computing time are greatly reduced. A lot of outstanding work on graph coloring problems has been achieved, and the theoretical lower bound of the chromatic number of random graph is one of them. Comparison between the theoretical lower bound and our computed approximate solution has been made for evaluation of proposed algorithm.

Location-based service with context data for a restaurant recommendation

Abstract: Utilizing Global Positioning System (GPS) technology, it is possible to find and recommend restaurants for users operating mobile devices. For recommending restaurants, Personal Digital Assistants or cellular phones only consider the location of restaurants. However, a user's background and environment information is assumed to be directly related to recommendation quality. In this paper, therefore, a recommender system using context information and a decision tree model for efficient recommendation is presented. This system considers location context, personal context, environment context, and user preference. Restaurant lists are obtained from location context, personal context, and environment context using the decision tree model. In addition, a weight value is used for reflecting user preferences. Finally, the system recommends appropriate restaurants to the mobile user. For this experiment, performance was verified using measurements such as k-fold cross-validation and Mean Absolute Error. As a result, the proposed system obtained an improvement in recommendation performance.

A design approach for Contextualized collaborative recommender system by nurturing the user interactive feedback

Abstract: The study nurtures the construction of user interaction in collaborative filtering system during recommendation time and thereby, seek for the contextual information to form contextualized collaborative recommendation. Firstly, contextualized token recommendation have been based on the memory based collaborative filtering techniques where recommendation for a given user is computed based on the weighted neighborhood technique for prediction which ultimately stores the long-term interest of the user and serves as a user context. Furthermore, our system let the users interact with the recommendation, which change the weight of the prediction and form the user short-term preference. We argue that user feedback upon recommendation given by a system, help a system to model user context during such feedback and in preliminary settings provide experimental evidence to support the claim.

Optimization of base stations positioning in mobile networks

Abstract: Cell planning is a very complex task because many aspects should be taken into account including the traffic distribution, geography, antenna height of base stations, frequency, and so forth. This becomes more complicated when several constraints are considered such as the quality of service, maximum output power of transmitters, types of power control, and other network requirements. The problems are assignment of mobile stations to the appropriate base station and finding good locations of the base stations. This paper presents a constraint satisfaction problem model for base station location in 3G W-CDMA uplink environments. This is similar to a warehouse location problem, so it is NP-hard and consequently cannot be practically solved by exact methods for real size networks. Thus we apply constraint satisfaction techniques such as variable ordering and value ordering to get good approximate solutions. The algorithm has been implemented by using both integer programming and constraint programming.

Mining large patterns with profit-based support in e-commerce

Abstract: In this paper, we propose an unique profit criterion as a new minimum support threshold for each item and exploit the criterion as multiple minimum supports in our algorithm. We then apply our profit-based association rule mining algorithm to generate large itemsets and show the result of our experiment. Experiment results carried on synthetic data set show that the proposed approach is efficient and effective in terms of reducing candidate itemsets and generating more profitable itemsets respectively.