• Journal of KIISE:Software and Applications
• /
• v.29 no.11
• /
• pp.785-794
• /
• 2002

Automatic stock trading systems should be able to solve various kinds of optimization problems such as market trend prediction, stock selection, and trading strategies, in a unified framework. But most of the previous trading systems based on supervised learning have a limit in the ultimate performance, because they are not mainly concerned in the integration of those subproblems. This paper proposes a stock trading system, called R-Trader, based on reinforcement teaming, regarding the process of stock price changes as Markov decision process (MDP). Reinforcement learning is suitable for Joint optimization of predictions and trading strategies. R-Trader adopts two popular reinforcement learning algorithms, temporal-difference (TD) and Q, for selecting stocks and optimizing other trading parameters respectively. Technical analysis is also adopted to devise the input features of the system and value functions are approximated by feedforward neural networks. Experimental results on the Korea stock market show that the proposed system outperforms the market average and also a simple trading system trained by supervised learning both in profit and risk management.

• Journal of KIISE:Computer Systems and Theory
• /
• v.33 no.10
• /
• pp.722-733
• /
• 2006

In scalable streaming application, there are two important knobs to tune to effectively exploit the underlying network resource and to maximize the user perceivable quality of service(QoS): layer selection and packet scheduling. In this work, we propose Semantics Aware Packet Scheduling (SAPS) algorithm to address these issues. Using packet dependency graph, SAPS algorithm selects a layer to maximize QoS. We aim at minimizing distortion in selecting layers. In inter-frame coded video streaming, minimizing packet loss does not imply maximizing QoS. In determining the packet transmission schedule, we exploit the fact that significance of each packet loss is different dependent upon its frame type and the position within group of picture(GOP). In SAPS algorithm, each packet is assigned a weight called QoS Impact Factor Transmission schedule is derived based upon weighted smoothing. In simulation experiment, we observed that QOS actually improves when packet loss becomes worse. The simulation results show that the SAPS not only maximizes user perceivable QoS but also minimizes resource requirements.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 1999.05a
• /
• pp.203-207
• /
• 1999

In this paper we propose the automated generation algorithm of penetration scenario using association mining technique. Until now known intrusion detections are classified into anomaly detection and misuse detection. The former uses statistical method, features selection, neural network method in order to decide intrusion, the latter uses conditional probability, expert system, state transition analysis, pattern matching for deciding intrusion. In proposed many intrusion detection algorithms unknown penetrations are created and updated by security experts. Our algorithm automatically generates penetration scenarios applying association mining technique to state transition technique. Association mining technique discovers efficient and useful unknown information in existing data. In this paper the algorithm we propose can automatically generate penetration scenarios to have been produced by security experts and is easy to cope with intrusions when it is compared to existing intrusion algorithms. Also It has advantage that maintenance cost is not high.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.32 no.3A
• /
• pp.297-309
• /
• 2007

We introduce two distributed IP address auto-configuration mechanisms for mobile ad hoc networks. RADA (Random ADdress Allocation) is based on random IP address selection, while LiA (Linear Address Allocation) assigns new addresses sequentially, using the current maximum IP address. An improved version of LiA, hewn as LiACR (Linear Address Allocation with Collision Resolution) further reduces the control overhead. Simulation results show that, when many nodes join a network during a short period, RADA assigns addresses more quickly than LiA and LiACR. However, RADA uses the address space less efficiently, due to its random allocation of IP addresses. Hence, RADA is particularly useful in battlefield scenarios or rescue operations where fast setup is needed, while LiA and LiACR are more suitable for ad hoc networks that are moderate, confined and subject to some form of governance control, such as that orchestrated by a wireless service provider.

• Journal of the Korea Society of Computer and Information
• /
• v.15 no.10
• /
• pp.113-120
• /
• 2010

Video proxy server which is located near clients can store the frequently requested video data in storage space in order to minimize initial latency and network traffic significantly. However, due to the limited storage space in video proxy server, an appropriate video selection method is needed to store the videos which are frequently requested by users. Thus, we present a time constraint multiple access technique to efficiently store the video in video proxy server. If the video is requested by user, it is temporarily stored during the predefined interval and then, delivered to the user. A video which is stored is deleted or moved into the storage space of video proxy server depending on the request condition. In addition, we propose a video deletion method in video proxy server for newly stored video data. The simulation results show that the proposed method performs better than other methods in terms of the block hit rate and the number of block deletion.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.14 no.6
• /
• pp.69-74
• /
• 2014

In this paper, an energy harvesting scheme is investigated in the multi-relay cooperation communication. Our proposal is to maximize the diversity gain and overcome the disadvantages of the limited power wireless devices. The best relay is selected based on the maximizing channel gains from source to relays so that the harvested energy at selected relay is best. If an internal power of the best relay is lower than a defined target power, the best relay will harvest the energy from the source, and the help of the source-destination link is changed to the second best relay. By this operation method, the diversity gain is maintained and the performance of the network is improved. Finally, performance of the proposed protocol is analyzed in terms of bit error rate, utilization efficiency, power collection efficiency.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2004.10a
• /
• pp.433-436
• /
• 2004

In this paper, we introduce a new Fuzzy Polynomial Neural Networks (FPNNS)-like structure whose neuron is based on the Fuzzy Set-based Fuzzy Inference System (FS-FIS) and is different from that of FPNNS based on the Fuzzy relation-based Fuzzy Inference System (FR-FIS) and discuss the ability of the new FPNNS-like structure named Fuzzy Set-based Polynomial Neural Networks (FSPNN). The premise parts of their fuzzy rules are not identical, while the consequent parts of the both Networks (such as FPNN and FSPNN) are identical. This difference results from the angle of a viewpoint of partition of input space of system. In other word, from a point of view of FS-FIS, the input variables are mutually independent under input space of system, while from a viewpoint of FR-FIS they are related each other. The proposed design procedure for networks architecture involves the selection of appropriate nodes with specific local characteristics such as the number of input variables, the order of the polynomial that is constant, linear, quadratic, or modified quadratic functions being viewed as the consequent part of fuzzy rules, and a collection of the specific subset of input variables. On the parameter optimization phase, we adopt Information Granulation (IC) based on HCM clustering algorithm and a standard least square method-based learning. Through the consecutive process of such structural and parametric optimization, an optimized and flexible fuzzy neural network is generated in a dynamic fashion. To evaluate the performance of the genetically optimized FSPNN (gFSPNN), the model is experimented with using the time series dataset of gas furnace process.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.13 no.2
• /
• pp.154-162
• /
• 2003

In this paper, Associative Memory with Cellular Neural Networks classifies and recognizes image patterns as an operator applied to image process. CNN processes nonlinear data in real-time like neural networks, and made by cell which communicates with each other directly through its neighbor cells as the Cellular Automata does. It is applied to the optimization problem, associative memory, pattern recognition, and computer vision. Image processing with CNN is appropriate to 2-D images, because each cell which corresponds to each pixel in the image is simultaneously processed in parallel. This paper shows the method for designing the structure of associative memory based on CNN and getting output image by choosing the most appropriate weight pattern among the whole learned weight pattern memories. Each template represents weight values between cells and updates them by learning. Hebbian rule is used for learning template weights and LMS algorithm is used for classification.

• IMMUNE NETWORK
• /
• v.2 no.4
• /
• pp.233-241
• /
• 2002

Background: Monoclonal antibodies (mAb) of rodent origin are produced with ease by hybridoma fusion technique, and have been successfully used as therapeutic reagents for humans after humanization by genetic engineering. However, utilization of these antibodies for therapeutic purpose has been limited by the fact that they act as immunogens in human body causing undesired side effects. So far, there have been several attempts to produce human mAbs for effective in vivo diagnostic or therapeutic reagents including the use of humanized xenomouse that is generated by mating knockout mice which lost Ig heavy and light chain genes by homologous recombination and transgenic mice having both human Ig heavy and light gene loci in their genome. Methods: Genomic DNA fragments of mouse Ig heavy and light chain were obtained from a mouse brain ${\lambda}$ genomic library by PCR screening and cloned into a targeting vector with ultimate goal of generating Ig knockout mouse. Results: Through PCR screening of the genomic library, three heavy chain and three light chain Ig gene fragments were identified, and restriction map of one of the heavy chain gene fragments was determined. Then heavy chain Ig gene fragments were subcloned into a targeting vector. The resulting construct was introduced into embryonic stem cells. Antibiotic selection of transfected cells is under the progress. Conclusion: Generation of xenomouse is particularly important in medical biotechnology. However, this goal is not easily achieved due to the technical difficulties as well as huge financial expenses. Although we are in the early stage of a long-term project, our results, at least, partially contribute the successful generation of humanized xenomouse in Korea.

• IMMUNE NETWORK
• /
• v.3 no.2
• /
• pp.89-95
• /
• 2003

Background: Unusually high amounts of N region addition and CDR3 length diversity were found in immunoglobulin (Ig) light chain Vk and Jk joins in patients with rheumatoid arthritis (RA). We sought to determine whether this finding is due to excessive activity of the enzyme responsible for N region addition (terminal deoxynucleotidyl transferase [TdT]) in B lineage cells in bone marrow or from positive antigenic selection of B cells with long CDR3 lengths. Methods: We used FACS to isolate $IgM^+/IgD^+$ B cells (predominantly naive) and $IgM^-/IgD^-$ B cells (predominantly class-switched) B cells from peripheral blood of a patient with RA known to have enrichment for long Vk CDR3s and from that of two normal controls. RT-PCR of VkIII transcripts was performed, followed by sequencing of individual cDNA clones. We analyzed the CDR3 lengths and N region additions in 97 clones. Results: There was enrichment for long CDR3 lengths (11 or 12 amino acids) in both $IgM^+/IgD^+$ and $IgM^-/IgD^-$ B cells in RA compared to B cell subsets in the normal controls. The $IgM^+/IgD^+$ B cell subset in RA was markedly enriched for N region addition and was similar to that seen in the $IgM^-/IgD^-$ subset. Conclusion: These data suggest that enrichment for N region addition and long CDR3 lengths in RA may result from unusually high or prolonged activity of TdT in bone marrow.