• Journal of Intelligence and Information Systems
• /
• v.16 no.4
• /
• pp.99-112
• /
• 2010

Ensemble learning is a method for improving the performance of classification and prediction algorithms. It is a method for finding a highly accurateclassifier on the training set by constructing and combining an ensemble of weak classifiers, each of which needs only to be moderately accurate on the training set. Ensemble learning has received considerable attention from machine learning and artificial intelligence fields because of its remarkable performance improvement and flexible integration with the traditional learning algorithms such as decision tree (DT), neural networks (NN), and SVM, etc. In those researches, all of DT ensemble studies have demonstrated impressive improvements in the generalization behavior of DT, while NN and SVM ensemble studies have not shown remarkable performance as shown in DT ensembles. Recently, several works have reported that the performance of ensemble can be degraded where multiple classifiers of an ensemble are highly correlated with, and thereby result in multicollinearity problem, which leads to performance degradation of the ensemble. They have also proposed the differentiated learning strategies to cope with performance degradation problem. Hansen and Salamon (1990) insisted that it is necessary and sufficient for the performance enhancement of an ensemble that the ensemble should contain diverse classifiers. Breiman (1996) explored that ensemble learning can increase the performance of unstable learning algorithms, but does not show remarkable performance improvement on stable learning algorithms. Unstable learning algorithms such as decision tree learners are sensitive to the change of the training data, and thus small changes in the training data can yield large changes in the generated classifiers. Therefore, ensemble with unstable learning algorithms can guarantee some diversity among the classifiers. To the contrary, stable learning algorithms such as NN and SVM generate similar classifiers in spite of small changes of the training data, and thus the correlation among the resulting classifiers is very high. This high correlation results in multicollinearity problem, which leads to performance degradation of the ensemble. Kim,s work (2009) showedthe performance comparison in bankruptcy prediction on Korea firms using tradition prediction algorithms such as NN, DT, and SVM. It reports that stable learning algorithms such as NN and SVM have higher predictability than the unstable DT. Meanwhile, with respect to their ensemble learning, DT ensemble shows the more improved performance than NN and SVM ensemble. Further analysis with variance inflation factor (VIF) analysis empirically proves that performance degradation of ensemble is due to multicollinearity problem. It also proposes that optimization of ensemble is needed to cope with such a problem. This paper proposes a hybrid system for coverage optimization of NN ensemble (CO-NN) in order to improve the performance of NN ensemble. Coverage optimization is a technique of choosing a sub-ensemble from an original ensemble to guarantee the diversity of classifiers in coverage optimization process. CO-NN uses GA which has been widely used for various optimization problems to deal with the coverage optimization problem. The GA chromosomes for the coverage optimization are encoded into binary strings, each bit of which indicates individual classifier. The fitness function is defined as maximization of error reduction and a constraint of variance inflation factor (VIF), which is one of the generally used methods to measure multicollinearity, is added to insure the diversity of classifiers by removing high correlation among the classifiers. We use Microsoft Excel and the GAs software package called Evolver. Experiments on company failure prediction have shown that CO-NN is effectively applied in the stable performance enhancement of NNensembles through the choice of classifiers by considering the correlations of the ensemble. The classifiers which have the potential multicollinearity problem are removed by the coverage optimization process of CO-NN and thereby CO-NN has shown higher performance than a single NN classifier and NN ensemble at 1% significance level, and DT ensemble at 5% significance level. However, there remain further research issues. First, decision optimization process to find optimal combination function should be considered in further research. Secondly, various learning strategies to deal with data noise should be introduced in more advanced further researches in the future.

• Journal of Broadcast Engineering
• /
• v.9 no.1
• /
• pp.17-31
• /
• 2004

This paper proposes an algorithm for the robust transmission of video in error Prone environment using multiple description codingby optimal split of DCT coefficients and rate-distortionoptimization framework. In MDC, a source signal is split Into several coded streams, which is called descriptions, and each description is transmitted to the decoder through different channel. Between descriptions, structured correlations are introduced at the encoder, and the decoder exploits this correlation to reconstruct the original signal even if some descriptions are missing. It has been shown that the MDC is more resilient than the singe description coding(SDC) against severe packet loss ratecondition. But the excessive redundancy in MDC, i.e., the correlation between the descriptions, degrades the RD performance under low PLR condition. To overcome this Problem of MDC, we propose a hybrid MDC method that controls the SDC/MDC switching according to channel condition. For example, the SDC is used for coding efficiency at low PLR condition and the MDC is used for the error resilience at high PLR condition. To control the SDC/MDC switching in the optimal way, RD optimization framework are used. Lagrange optimization technique minimizes the RD-based cost function, D+M, where R is the actually coded bit rate and D is the estimated distortion. The recursive optimal pet-pixel estimatetechnique is adopted to estimate accurate the decoder distortion. Experimental results show that the proposed optimal split of DCT coefficients and SD/MD switching algorithm is more effective than the conventional MU algorithms in low PLR conditions as well as In high PLR condition.

• Journal of KIISE:Databases
• /
• v.29 no.1
• /
• pp.79-88
• /
• 2002

Data on the Internet are usually represented and transfered as XML. the XML data is represented as a tree and therefore, object repositories are well-suited to store and query them due to their modeling power. XML queries are represented as regular path expressions and evaluated by traversing each object of the tree in object repositories. Several indexes are proposed to fast evaluate regular path expressions. However, in some cases they may not cover all possible paths because they require a great amount of disk space. In order to efficiently evaluate the queries in such cases, we propose an optimized traversing which combines the signature method and block traversing. The signature approach shrink the search space by using the signature information attached to each object, which hints the existence of a certain label in the sub-tree. The block traversing reduces disk I/O by early evaluating the reachable objects in a page. We conducted diverse experiments to show that the hybrid approach achieves a better performance than the other naive ones.

• Journal of the Korean Solar Energy Society
• /
• v.27 no.2
• /
• pp.19-27
• /
• 2007

The previous study was conducted to develop an air source multi heat pump system that could be operated with the solar collector and air source heat exchangers as heat source of the system. There is a winter-sowing problems in air source multi heat pump system when the outdoor temperature goes down under freezing point. The winter-sowing problem was solved by adapting R-22 refrigerant as working fluid in the previous study. However, when the system operated at high temperature, another problems are come out such as overheating of the solar collector outlet which lead to the superheat of the compressor inlet of the heat pump system. The condition could deteriorates a compressor in some case. In this study, we installed the anti-superheating devices on the previously developed system. As results of system performance test, COP of the system with anti-superheating technique is 2.4. It is a little improved COP compare to previous study's 2.23. In the results of multi heat source heating system, during operating solar collector, COP is relatively high between $200\;W/m^2$ and $400\;W/m^2$ solar intensity. It is recommended to extend the study on performance optimization with balancing the solar collect and capacity of compressor at higher solar irradiation conditions.

• Korean Journal of Optics and Photonics
• /
• v.22 no.6
• /
• pp.247-255
• /
• 2011

The optical performance of LLF(Lenticular Lens Film)-based backlight was studied by using optical simulation as functions of the aspect ratio and the refractive index of lenticular lenses. In order to perform reliable simulation, the BSDF(bi-directional scattering distribution function) of the scattering dots on the bottom surface of the light guide plate was obtained as a superposition of the Lambertian and the elliptic Gaussian distribution components by comparing the experimental results and the simulation for the luminance distribution on the light guide. Based on this approach, an appropriate BSDF of the scattering dots of the light guide was constructed. The resultant values of the optimized aspect ratio and the refractive were found to be 1.25 and 1.65, respectively. In spite of the hybrid aspects of LLF incorporating both diffusing and collimating functions, the optical performance, in particular the on-axis luminance of LLF-based backlight was inferior by about 20% compared to that of conventional backlights adopting one prism film. However, the combination of two lenticular lens films resulted in comparable luminance gain as well as smooth decrease in the luminance with the viewing angle without exhibiting any side lobes.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.10a
• /
• pp.243-245
• /
• 2021

Existing beacons are suitable for providing untact services, but they have the disadvantage of difficulty in accurate indoor positioning because the deviation in signal strength increases depending on the environment. In general, trilateration technique can reduce deviation, but if the distance between beacons is quite irregular, it becomes difficult to apply the algorithm. Therefore, in this paper, we studied how to reduce the signal power measurement error between beacons. First, we transformed the distance measurement formula using RSSI, assuming that the TX values were the same. In addition, we compared measurement errors with existing beacons by searching beacons with beacons scanner applications implemented with Android. As a result, it was confirmed that if a certain distance was further away, the measurement was measured more accurately than the non-changing beacon. Through this, accurate indoor positioning will be possible even in various disability situations. It is also expected that there will be more cases of establishing services that combine beacon with non-face-to-face services.

• International Journal of Computer Science & Network Security
• /
• v.23 no.11
• /
• pp.99-109
• /
• 2023

The darknet is frequently referred to as the hub of illicit online activity. In order to keep track of real-time applications and activities taking place on Darknet, traffic on that network must be analysed. It is without a doubt important to recognise network traffic tied to an unused Internet address in order to spot and investigate malicious online activity. Any observed network traffic is the result of mis-configuration from faked source addresses and another methods that monitor the unused space address because there are no genuine devices or hosts in an unused address block. Digital systems can now detect and identify darknet activity on their own thanks to recent advances in artificial intelligence. In this paper, offer a generalised method for deep learning-based detection and classification of darknet traffic. Furthermore, analyse a cutting-edge complicated dataset that contains a lot of information about darknet traffic. Next, examine various feature selection strategies to choose a best attribute for detecting and classifying darknet traffic. For the purpose of identifying threats using network properties acquired from darknet traffic, devised a hybrid deep learning (DL) approach that combines Recurrent Neural Network (RNN) and Bidirectional LSTM (BiLSTM). This probing technique can tell malicious traffic from legitimate traffic. The results show that the suggested strategy works better than the existing ways by producing the highest level of accuracy for categorising darknet traffic using the Black widow optimization algorithm as a feature selection approach and RNN-BiLSTM as a recognition model.

• Journal of Korean Society of Transportation
• /
• v.22 no.2 s.73
• /
• pp.145-154
• /
• 2004

The growing demand for customer-responsive, made-to-order manufacturing is stimulating the need for improved dynamic decision-making processes in commercial fleet operations. Moreover, the rapid growth of electronic commerce through the internet is also requiring advanced and precise real-time operation of vehicle fleets. Accompanying these demand side developments/pressures, the growing availability of technologies such as AVL(Automatic Vehicle Location) systems and continuous two-way communication devices is driving developments on the supply side. These technologies enable the dispatcher to identify the current location of trucks and to communicate with drivers in real time affording the carrier fleet dispatcher the opportunity to dynamically respond to changes in demand, driver and vehicle availability, as well as traffic network conditions. This research investigates key aspects of real time dynamic routing and scheduling problems in fleet operation particularly in a truckload pickup-and-delivery problem under various settings, in which information of stochastic demands is revealed on a continuous basis, i.e., as the scheduled routes are executed. The most promising solution strategies for dealing with this real-time problem are analyzed and integrated. Furthermore, this research develops. analyzes, and implements hybrid algorithms for solving them, which combine fast local heuristic approach with an optimization-based approach. In addition, various partitioning algorithms being able to deal with large fleet of vehicles are developed based on 'divided & conquer' technique. Simulation experiments are developed and conducted to evaluate the performance of these algorithms.

• Composites Research
• /
• v.34 no.1
• /
• pp.35-46
• /
• 2021

This paper aims to reduce weight by replacing the reinforcements of the B-pillar used in vehicles with CFRP(Carbon Fiber Reinforced Plastics) and GFRP(Glass Fiber Reinforced Plastics) from the existing steel materials. For this, it is necessary to secure structural stability that can replace the existing B-pillar while reducing the weight. Existing B-pillar are composed of steel reinforcements of various shapes, including a steel outer. Among these steel reinforcements, two steel reinforcements are to be replaced with composite materials. Each steel reinforcement is manufactured separately and bonded to the B-pillar outer by welding. However, the composite reinforcements presented in this paper are manufactured at once through compression and injection processes using patch-type CFRP and rib-structured GFRP. CFRP is attached to the high-strength part of the B-pillar to resist side loads, and the GFRP ribs are designed to resist torsion and side loads through a topology optimization technique. Through structural analysis, the designed composite B-pillar was compared with the existing B-pillar, and the weight reduction ratio was calculated.

• Journal of Korean Society of Transportation
• /
• v.23 no.3 s.81
• /
• pp.35-47
• /
• 2005

This paper is to review a literature concerning Bus Transit Route Network Design(BTRNDP), to describe a future study direction for a systematic application for the BTRNDP. Since a bus transit uses a fixed route, schedule, stop, therefore an approach methodology is different from that of auto network design problem. An approach methodology for BTRNDP is classified by 8 categories: manual & guideline, market analysis, system analytic model. heuristic model. hybrid model. experienced-based model. simulation-based model. mathematical optimization model. In most previous BTRNDP, objective function is to minimize user and operator costs, and constraints on the total operator cost, fleet size and service frequency are common to several previous approach. Transit trip assignment mostly use multi-path trip assignment. Since the search for optimal solution from a large search space of BTRNDP made up by all possible solutions, the mixed combinatorial problem are usually NP-hard. Therefore, previous researches for the BTRNDP use a sequential design process, which is composed of several design steps as follows: the generation of a candidate route set, the route analysis and evaluation process, the selection process of a optimal route set Future study will focus on a development of detailed OD trip table based on bus stop, systematic transit route network evaluation model. updated transit trip assignment technique and advanced solution search algorithm for BTRNDP.