• Journal of Intelligence and Information Systems
• /
• v.24 no.2
• /
• pp.221-241
• /
• 2018

Deep learning is getting attention recently. The deep learning technique which had been applied in competitions of the International Conference on Image Recognition Technology(ILSVR) and AlphaGo is Convolution Neural Network(CNN). CNN is characterized in that the input image is divided into small sections to recognize the partial features and combine them to recognize as a whole. Deep learning technologies are expected to bring a lot of changes in our lives, but until now, its applications have been limited to image recognition and natural language processing. The use of deep learning techniques for business problems is still an early research stage. If their performance is proved, they can be applied to traditional business problems such as future marketing response prediction, fraud transaction detection, bankruptcy prediction, and so on. So, it is a very meaningful experiment to diagnose the possibility of solving business problems using deep learning technologies based on the case of online shopping companies which have big data, are relatively easy to identify customer behavior and has high utilization values. Especially, in online shopping companies, the competition environment is rapidly changing and becoming more intense. Therefore, analysis of customer behavior for maximizing profit is becoming more and more important for online shopping companies. In this study, we propose 'CNN model of Heterogeneous Information Integration' using CNN as a way to improve the predictive power of customer behavior in online shopping enterprises. In order to propose a model that optimizes the performance, which is a model that learns from the convolution neural network of the multi-layer perceptron structure by combining structured and unstructured information, this model uses 'heterogeneous information integration', 'unstructured information vector conversion', 'multi-layer perceptron design', and evaluate the performance of each architecture, and confirm the proposed model based on the results. In addition, the target variables for predicting customer behavior are defined as six binary classification problems: re-purchaser, churn, frequent shopper, frequent refund shopper, high amount shopper, high discount shopper. In order to verify the usefulness of the proposed model, we conducted experiments using actual data of domestic specific online shopping company. This experiment uses actual transactions, customers, and VOC data of specific online shopping company in Korea. Data extraction criteria are defined for 47,947 customers who registered at least one VOC in January 2011 (1 month). The customer profiles of these customers, as well as a total of 19 months of trading data from September 2010 to March 2012, and VOCs posted for a month are used. The experiment of this study is divided into two stages. In the first step, we evaluate three architectures that affect the performance of the proposed model and select optimal parameters. We evaluate the performance with the proposed model. Experimental results show that the proposed model, which combines both structured and unstructured information, is superior compared to NBC(Naïve Bayes classification), SVM(Support vector machine), and ANN(Artificial neural network). Therefore, it is significant that the use of unstructured information contributes to predict customer behavior, and that CNN can be applied to solve business problems as well as image recognition and natural language processing problems. It can be confirmed through experiments that CNN is more effective in understanding and interpreting the meaning of context in text VOC data. And it is significant that the empirical research based on the actual data of the e-commerce company can extract very meaningful information from the VOC data written in the text format directly by the customer in the prediction of the customer behavior. Finally, through various experiments, it is possible to say that the proposed model provides useful information for the future research related to the parameter selection and its performance.

• Journal of Bio-Environment Control
• /
• v.31 no.4
• /
• pp.384-392
• /
• 2022

The objective of this study was to estimated nitrogen content and chlorophyll using RGB, Hyperspectral sensors to diagnose of nitrogen nutrition in apple tree leaves. Spectral data were acquired through image processing after shooting with high resolution RGB and hyperspectral sensor for two-year-old 'Hongro/M.9' apple. Growth data measured chlorophyll and leaf nitrogen content (LNC) immediately after shooting. The growth model was developed by using regression analysis (simple, multi, partial least squared) with growth data (chlorophyll, LNC) and spectral data (SPAD meter, color vegetation index, wavelength). As a result, chlorophyll and LNC showed a statistically significant difference according to nitrogen fertilizer level regardless of date. Leaf color became pale as the nutrients in the leaf were transferred to the fruit as over time. RGB sensor showed a statistically significant difference at the red wavelength regardless of the date. Also hyperspectral sensor showed a spectral difference depend on nitrogen fertilizer level for non-visible wavelength than visible wavelength at June 10th and July 14th. The estimation model performance of chlorophyll, LNC showed Partial least squared regression using hyperspectral data better than Simple and multiple linear regression using RGB data (Chlorophyll R²: 81%, LNC: 81%). The reason is that hyperspectral sensor has a narrow Full Half at Width Maximum (FWHM) and broad wavelength range (400-1,000 nm), so it is thought that the spectral analysis of crop was possible due to stress cause by nitrogen deficiency. In future study, it is thought that it will contribute to development of high quality and stable fruit production technology by diagnosis model of physiology and pest for all growth stage of tree using hyperspectral imagery.

• Korean Journal of Poultry Science
• /
• v.20 no.2
• /
• pp.107-114
• /
• 1993

The structure change in the broiler industry of Korea has been taking place very rapidly toward the direction of integrated production and marketing system. During the course of the evolution into the integrated structure, however, a number of small business entities, uncountable nationally but a minimum of a dozen, that previously engaged in the supply of production factors or live bird transaction also tended to get involved in the new system as a disguised or pseudo-integrator, having brought a lot of problems not only to the farmers but also to the development of integrated structure. The pseudo-integrator is generally characterized by limited functions in such a way of supplying chicks and feeds to and collecting grown birds from farmers under the contract at a fixed farmer's payment in practically the same pattern as a partial or quasi-integration, which intends to act as if an integrator in a stratagem to simply enjoy a margin simply from selling supplies and buying products for a certain period of time. The grower making a contract with appears to be a farmer who used to be an Independent and speculating but not able to join in the normal system of evolving integration. The problems of a pseudo-integrator center on the financial loss to a contract farmer, because the falsified integrator has to become easily bankrupt and run away when the price of live broiler continually stays below the cost of production, even though he is able to make a tremendous profit otherwise which is the real purpose of the operation for. It is true that the volatile market, fluctuating the price up more than doubled in a month and down to a half in a few weeks, makes the pseudo-integrator find the room for such a fraudulence. In addition, its activity also adversely affect the evolution of the integration due to rather negative image on structure change in general. It is recommended that the farmers need to better understand the real picture of the integrated system so as not to be swindled by a disguised, small scale agribusiness agent. By the same token, it is also equally required to have the whole industry integrated completely as early as possible. The Joint effort to get rid of pseudo-integrators' problems shall be put for the industry development moving toward the integration. No doubt a pseudo-integrator must be a temporary player for chance emerging during the course of structure change into the integrated, though.

• The Journal of the Korea Contents Association
• /
• v.14 no.4
• /
• pp.47-57
• /
• 2014

The horro movie which was released in 2003, was a box office hit drawing audiences totaling 3,140,000 people. Then, there were partial opinions that it was difficult to understand the content of the movie. However, it was most favorably received in the horror movies which were released before and after 2003. The 4 main characters in the movie are a family. The family conflicts begin as the dad has an affair. Muhyeon(the dad)'s infidelity leads his wife to kill herself. The suicide of the wife and mom again leads Suyeon(the younger sister) to be dead. At this moment, Suyeon's death is related to Eunju(the step mom). Sumi(the older sister) is deeply hurt by the fact that Eunju left her younger sister's death after she was dead. But at the same time, she realizes that she is also related to Eunju's act. And she felt guilty about her younger sister's death. Eunju was also hurt by Sumi's biting remarks. However, she feels guilty about the fact that she made Suyeon be dead. Director Jiun Kim said in the interview when it was released, "this movie is about hurts and guilt among families." These hurts and guilt are expressed by horror images that imagination and reality are mixed. Hurts are expressed by virtual images and guilt is done by real images. This paper presented still pictures which explain the virtual and real images and examined what hurts and guilt mean. And it described the symbolic meanings of three media which cause conflicts among main characters.

• The KIPS Transactions:PartB
• /
• v.9B no.3
• /
• pp.359-368
• /
• 2002

In this paper, we propose effective boundary line extraction algorithm for moving objects by matching error image and moving vectors, and fast tracking algorithm for moving object by partial boundary lines. We extracted boundary line for moving object by generating seeds with probability distribution function based on Watershed algorithm, and by extracting boundary line for moving objects through extending seeds, and then by using moving vectors. We processed tracking algorithm for moving object by using a part of boundary lines as features. We set up a part of every-direction boundary line for moving object as the initial feature vectors for moving objects. Then, we tracked moving object within current frames by using feature vector for the previous frames. As the result of the simulation for tracking moving object on the real images, we found that tracking processing of the proposed algorithm was simple due to tracking boundary line only for moving object as a feature, in contrast to the traditional tracking algorithm for active contour line that have varying processing cost with the length of boundary line. The operations was reduced about 39% as contrasted with the full search BMA. Tracking error was less than 4 pixel when the feature vector was $(15\times{5)}$ through the information of every-direction boundary line. The proposed algorithm just needed 200 times of search operation.

• Economic and Environmental Geology
• /
• v.28 no.3
• /
• pp.199-211
• /
• 1995

The deep sea camera system could render it possible to obtain the detailed information of the nodule distribution, but difficult to estimate nodule abundance quantitatively. In order to estimate nodule abundance quantitatively from deep seabed photographs, the nodule abundance equation was derived from the box core data obtained in KODOS area(long.: $154^{\circ}{\sim}151^{\circ}W$, lat.: $9^{\circ}{\sim}12^{\circ}N$) during two survey cruises carried out in 1989 and 1990. The regression equation derived by considering extent of burial of nodule to Handa's equation compensates for the abundance error attributable to partial burial of some nodules by sediments. An average long axis and average extent of burial of nodules in photographed area are determined according to the surface textures of nodules, and nodule coverage is calculated by the image analysis method. Average nodule abundance estimated from seabed photographs by using the equation is approximately 92% of the actual average abundance in KODOS area. The measured sampling points by box core or free fall grab are in general very sparse and hence nodule abundance distribution should be interpolated and extrapolated from measured data to uncharacterized areas. The another goal of this study is to depict continuous distribution of nodule abundance in KODOS area by using PC-version of geostatistical model in which several stages are systematically proceeded. Geostatistics was used to analyse spatial structure and distribution of regionalized variable(nodule abundance) within sets of real data. In order to investigate the spatial structure of nodule abundance in KODOS area, experimental variograms were calculated and fitted to a spherical models in isotropy and anisotropy, respectively. The spherical structure models were used to map out distribution of the nodule abundance for isotropic and anisotropic models by using the kriging method. The result from anisotropic model is much more reliable than one of isotropic model. Distribution map of nodule abundance produced by PC-version of geostatistical model indicates that approximately 40% of KODOS area is considered to be promising area(nodule abundance > $5kg/m^2$) for mining in case of anisotropy.

• Journal of Biosystems Engineering
• /
• v.43 no.2
• /
• pp.148-159
• /
• 2018

Purpose: Unmanned air vehicle (UAV) remote sensing was applied to test various vegetation indices and make prediction models of protein content of rice for monitoring grain quality and proper management practice. Methods: Image acquisition was carried out by using NIR (Green, Red, NIR), RGB and RE (Blue, Green, Red-edge) camera mounted on UAV. Sampling was done synchronously at the geo-referenced points and GPS locations were recorded. Paddy samples were air-dried to 15% moisture content, and then dehulled and milled to 92% milling yield and measured the protein content by near-infrared spectroscopy. Results: Artificial neural network showed the better performance with $R^2$ (coefficient of determination) of 0.740, NSE (Nash-Sutcliffe model efficiency coefficient) of 0.733 and RMSE (root mean square error) of 0.187% considering all 54 samples than the models developed by PR (polynomial regression), SLR (simple linear regression), and PLSR (partial least square regression). PLSR calibration models showed almost similar result with PR as 0.663 ($R^2$) and 0.169% (RMSE) for cloud-free samples and 0.491 ($R^2$) and 0.217% (RMSE) for cloud-shadowed samples. However, the validation models performed poorly. This study revealed that there is a highly significant correlation between NDVI (normalized difference vegetation index) and protein content in rice. For the cloud-free samples, the SLR models showed $R^2=0.553$ and RMSE = 0.210%, and for cloud-shadowed samples showed 0.479 as $R^2$ and 0.225% as RMSE respectively. Conclusion: There is a significant correlation between spectral bands and grain protein content. Artificial neural networks have the strong advantages to fit the nonlinear problem when a sigmoid activation function is used in the hidden layer. Quantitatively, the neural network model obtained a higher precision result with a mean absolute relative error (MARE) of 2.18% and root mean square error (RMSE) of 0.187%.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.19 no.5
• /
• pp.184-189
• /
• 2018

Recently, research on lightweight automobiles is increasing continuously to respond to the high safety standards and environmental regulations. The application of high strength steel is one of the effective methods for developing lightweight car bodies. A TWB (Tailor Welded Blank) is major method that allows partial high strength with light weighting using a multi-thickness and multi-material welded blank. On the other hand, additional welding process is required to prepare the blank and quality control for the welding process also required. To secure this point, the TRB (Tailor Rolled Blank) method was suggested. In the TRB method, the thickness of sheet is controlled by the rolling presses and the production efficiency is much higher than welding in TWB methods. In this study, the formability of high strength TRB steel plate was analyzed to examine the rolling effect of the blank. The formability of the specimen was tested using 0.8 and 1 mm thick steel sheets for the TRB plate. The strain was analyzed by the digital image sensing of grid markings on the specimen and the forming limit diagram was calculated. An Erichsen test for the 0.8 and 1 mm thick TRB specimens was carried out and the formability was investigated by comparing with FE analysis.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.18 no.7
• /
• pp.249-256
• /
• 2017

Luxury goods products have a steady stream of consumers, and the latest trend is the sell and purchase markets. The consumers age groups also vary widely. This research examined the mediating effect of self-construal of the relationship between brand consistency and intention to purchase luxury fashion products. The analysis classified the brand consistency of luxury fashion products as social, actual, and the deal. First, the order of consistency that had an effect on self-construal appeared in the order of social cohesion, actual coincidence, and ideal coincidence. Second, an analysis of mediating effect of self-construal showed that only self-construal appeared to have a partial mediating effect on the relationship between the intention to purchase and the actual and ideal consistency. Such result can be interpreted as that when purchasing luxury fashion products, consumers have a tendency to desire that the brand would express their personal image and that it would match the perception they had of the brand, rather than a tendency to perceive the brand as a tool to help them perform their desired social role. These results suggest that consumers who consume the brand desire to have their personalities and subjective environmental conditions well expressed.

• The KIPS Transactions:PartB
• /
• v.9B no.2
• /
• pp.245-254
• /
• 2002

IC package marking inspection system using machine vision locates and identifies the target elements from input image, and decides the quality of marking by comparing the extracted target elements with the standard patterns. This paper proposes an adaptive multi-level thresholding (AMLT) method which is suitable for a series of operations such as locating the target IC package, extracting the characters, and detecting the Pinl dimple. It also proposes a dynamic matching unit selection (DMUS) method which is robust to noises as well as effective to catch out the local marking errors. The main idea of the AMLT method is to restrict the inputs of Otsu's thresholding algorithm within a specified area and a partial range of gray values. Doing so, it can adapt to the specific domain. The DMUS method dynamically selects the matching unit according to the result of character extraction and layout analysis. Therefore, in spite of the various erroneous situation occurred in the process of character extraction and layout analysis, it can select minimal matching unit in any environment. In an experiment with 280 IC package images of eight types, the correct extracting rate of IC package and Pinl dimple was 100% and the correct decision rate of marking quality was 98.8%. This result shows that the proposed methods are effective to IC package marking inspection.