• Journal of the Korean Data and Information Science Society
• /
• v.27 no.4
• /
• pp.911-923
• /
• 2016

Machine learning becomes increasingly important in the big data era. Logistic regression is a type of classification in machine leaning, and has been widely used in various fields, including medicine, economics, marketing, and social sciences. Rhipe that integrates R and Hadoop environment, has not been discussed by many researchers owing to the difficulty of its installation and MapReduce implementation. In this paper, we present the MapReduce implementation of Gradient Descent algorithm and Newton-Raphson algorithm for logistic regression using Rhipe. The Newton-Raphson algorithm does not require a learning rate, while Gradient Descent algorithm needs to manually pick a learning rate. We choose the learning rate by performing the mixed procedure of grid search and binary search for processing big data efficiently. In the performance study, our Newton-Raphson algorithm outpeforms Gradient Descent algorithm in all the tested data.

• The KIPS Transactions:PartA
• /
• v.13A no.6 s.103
• /
• pp.481-488
• /
• 2006

FRP(Functional Reactive Programming) is a kind of embedded language in Haskell, it declaratively program reactive system based on two essential high-order types named behavior and events. This Paper design and implementation RT-FRP(Real-time Functional Reactive Programming) translator for using FRP in embedded systems with many constraints. The RT-FRP translator generates a C Program from an RT-FRP program according to the operational semantics of the RT-FRP language. To show the effectiveness of the RT-FRP translator, we loaded and executed the test program generated by the translator onto a real embedded system, LEGO Mindstorm. According to the experimental result, the reactive system software can be programmed more concisely using RT-FRP than using an imperative counter part although the size of the binary code is rather increased.

• The KIPS Transactions:PartA
• /
• v.11A no.2
• /
• pp.109-114
• /
• 2004

This paper presents a high-speed bit-parallel systolic divider for computing modular division A($\chi$)/B($\chi$) mod G($\chi$) in finite fields GF$(2^m)$. The presented divider is based on the binary GCD algorithm and verified through FPGA implementation. The proposed architecture produces division results at a rate of one every 1 clock cycles after an initial delay of 5m-2. Analysis shows that the proposed divider provides a significant reduction in both chip area and computational delay time compared to previously proposed systolic dividers with the same I/O format. In addition, since the proposed architecture does not restrict the choice of irreducible polynomials and has regularity and modularity, it provides a high flexibility and Scalability with respect to the field size m. Therefore, the proposed divider is well suited to VLSI implementation.

• Journal of Digital Convergence
• /
• v.11 no.9
• /
• pp.311-318
• /
• 2013

For face detection techniques, the correctness of detection decreases with different lightings and backgrounds so such requires new methods and techniques. This study has aimed to obtain data for reasoning human emotional information by analyzing the components of the eyes and mouth that are critical in expressing emotions. To do this, existing problems in detecting face are addressed and a detection method that has a high detection rate and fast processing speed good at detecting environmental elements is proposed. This method must detect a specific part (eyes and a mouth) by using Haar-like Feature technique with the application of an integral image. After which, binaries detect elements based on color information, dividing the face zone and skin zone. To generate correct shape, the shape of detected elements is generated by using a bezier curve-a curve generation algorithm. To evaluate the performance of the proposed method, an experiment was conducted by using data in the Face Recognition Homepage. The result showed that Haar-like technique and bezier curve method were able to detect face elements more elaborately.

• Journal of the Korean Institute of Telematics and Electronics A
• /
• v.30A no.3
• /
• pp.1-15
• /
• 1993

In this paper, we propose over-reporting avoidance scheme which avoids congestion of network traffics by adjusting managed system's over-reporting, on the OSI network management model which reports events from managed system to managing system. In case of reporting events from managed system to managing system, management traffic concentration occurs, and it causes over-loading on the managing system and congestion on the network. This scheme takes advantage of feedback from managing system to managed system. Managed system transmits event reports as much as maximum event pertime allocated to itself to managing system, and it sets it's management variables to LOCK state and stops event reports as Threshold time is reached. At the time, managing system directs event reports again by using M-set primitive with referring it's status. With this scheme, distributed processing, dynamic network adaptation, convergence of optimal operation point is possible. In addition to it, a fairness is assured. In order to detect characteristics of the Unary feedback over-reporting avoidance scheme. It is observed a control capability of the event reporting and fairness of each nodes through measuring. ThresholdTime value. It is measured a number of mean activating nodes and maintained time of LOCK state according to event reporting load, and also measured lost ratio of management packet, queuing delay in managing system, and goodput to observe effects of general packet load. Binary feedback scheme. Unary feedback overreporting avoidance scheme and raw scheme on the OSI network management system each are compared and analyzed, and finally proved that the scheme proposed in this study performs better.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.16 no.2
• /
• pp.179-184
• /
• 2006

An intuitive understanding of the dynamic pattern generation in asymmetric networks may be considered an essential component in developing models for the dynamic information processing. It has been reported that the neural network with cyclic connections generates multiple limit cycles. The dynamics of discrete time network with cyclic connections has been investigated intensively. However, the dynamics of a cyclic connection neural network in continuous-time has not been well-known due to the considerable complexity involved in its calculation. In this paper, the dynamic behavior of a continuous-time cyclic connection neural network, in which each neuron is connected only to its nearest neurons with binary synaptic weights of ${\pm}1$, has been investigated. Furthermore, the dynamics and stability of the network have been analyzed using a piece-wise linear approximation.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.21 no.6
• /
• pp.705-711
• /
• 2011

This paper proposes an effective algorithm of license plate recognition for industrial applications. By applying Canny edge detection on a vehicle image, it is possible to find a connected rectangular, which is a strong candidate for license plate. The color information of license plate separates plates into white and green. Then, OTSU binary image processing and foreground neighbor pixel propagation algorithm CLNF will be applied to each license plates to reduce noise except numbers and letters. Finally, through labeling, numbers and letters will be extracted from the license plate. Letter and number regions, separated from the plate, pass through mesh method and thinning process for extracting feature vectors by X-Y projection method. The extracted feature vectors are compared with the pre-learned weighting values by backpropagation neural network to execute final recognition process. The experiment results show that the proposed license plate recognition algorithm works effectively.

• Korean Chemical Engineering Research
• /
• v.48 no.2
• /
• pp.205-211
• /
• 2010

The water-sewage cohesion agent(polyaluminumchloride(PAC)) and NaOH were used to synthesize $Al(OH)_3$. For various additions of NaOH, characteristics of the synthesized $Al(OH)_3$ was analysed by XRD, SEM and PSA. According to XRD analysis, small amount of NaOH(NaOH:PAC=15g:100g) resulted in amorphous form of $Al(OH)_3$. By increasing NaOH(NaOH:PAC=20g:100g), the mixture of gibbsite(37%), bayerite(35%) and boehmite(28%) were produced. By adding more NaOH(NaOH:PAC=25g:100g), binary mixtures of gibbsite(67%) and bayerite(33%) were formed. Finally, high addition of NaOH(NaOH:PAC=30g:100g) gave the high concentration of gibbsite(gibbsite:bayerite=83:17). Also, SEM analysis indicated that the product featured the plate form with 20 and 30g of NaOH addition. Furthermore it was found that the particle size of the product decreased with the addition of NaOH.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.34 no.1C
• /
• pp.72-80
• /
• 2009

Look-up table based inverse halftoning from the digital color halftone image is proposed in this paper, which uses vector adaptive filter for the nonexisting patterns in the table. Halftone image is obtained from a continuous -tone image, which can be restored into continuous one from the digital binary image by way of inverse halftoning method. Look-up table based method usually processes fast and has even performances over the various halftoning. The numbers of pixels in the pattern of table and the method how to define the table elements for each R, G, B channels can effect largely for its performance. The proposed method uses 16 pixels in the table considering the diversity of the expressions from their patterns and with memory size as well. This also proposed how to combine R, G, B channels into one. Experimental results showed the better performance in the expression of colors, better color restoration and the short processing time compared with the conventional ones.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.13 no.3
• /
• pp.594-601
• /
• 2009

Shooting simulation systems not only reduce a great amount of expense and time for military exercises but also prevent accidents. In particular, the shooting simulation systems using laser beam have an advantage which is very similar to the shooting exercise that uses real bullets. However, real time technique for laser beam recognition in a target image is necessary. The method proposed in this paper takes a difference image from two adjacent image frames. Then a thresholding is applied on this difference image to discriminate laser beam from background. To decide the threshold value the intensity distribution of background points is modeled assuming normal distribution. Then a noise reduction and a region segmentation are applied on the binary image to find the position of a laser beam. The time complexity of this process depends on the size of an image multiplied by the size of a mask used in the noise reduction process. The experimental result showed that the accuracy of the system was 93.3%. Even in the inaccurate cases the beam was always found in the resultant region.