• The Journal of Bigdata
• /
• v.4 no.2
• /
• pp.1-12
• /
• 2019

As Artificial Intelligence (AI) technology develops, it is applied to various fields such as image, voice, and text. AI has shown fine results in certain areas. Researchers have tried to predict the stock market by utilizing artificial intelligence as well. Predicting the stock market is known as one of the difficult problems since the stock market is affected by various factors such as economy and politics. In the field of AI, there are attempts to predict the ups and downs of stock price by studying stock price patterns using various machine learning techniques. This study suggest a way of predicting stock price patterns based on the Convolutional Neural Network(CNN) among machine learning techniques. CNN uses neural networks to classify images by extracting features from images through convolutional layers. Therefore, this study tries to classify candlestick images made by stock data in order to predict patterns. This study has two objectives. The first one referred as Case 1 is to predict the patterns with the images made by the same-day stock price data. The second one referred as Case 2 is to predict the next day stock price patterns with the images produced by the daily stock price data. In Case 1, data augmentation methods - random modification and Gaussian noise - are applied to generate more training data, and the generated images are put into the model to fit. Given that deep learning requires a large amount of data, this study suggests a method of data augmentation for candlestick images. Also, this study compares the accuracies of the images with Gaussian noise and different classification problems. All data in this study is collected through OpenAPI provided by DaiShin Securities. Case 1 has five different labels depending on patterns. The patterns are up with up closing, up with down closing, down with up closing, down with down closing, and staying. The images in Case 1 are created by removing the last candle(-1candle), the last two candles(-2candles), and the last three candles(-3candles) from 60 minutes, 30 minutes, 10 minutes, and 5 minutes candle charts. 60 minutes candle chart means one candle in the image has 60 minutes of information containing an open price, high price, low price, close price. Case 2 has two labels that are up and down. This study for Case 2 has generated for 60 minutes, 30 minutes, 10 minutes, and 5minutes candle charts without removing any candle. Considering the stock data, moving the candles in the images is suggested, instead of existing data augmentation techniques. How much the candles are moved is defined as the modified value. The average difference of closing prices between candles was 0.0029. Therefore, in this study, 0.003, 0.002, 0.001, 0.00025 are used for the modified value. The number of images was doubled after data augmentation. When it comes to Gaussian Noise, the mean value was 0, and the value of variance was 0.01. For both Case 1 and Case 2, the model is based on VGG-Net16 that has 16 layers. As a result, 10 minutes -1candle showed the best accuracy among 60 minutes, 30 minutes, 10 minutes, 5minutes candle charts. Thus, 10 minutes images were utilized for the rest of the experiment in Case 1. The three candles removed from the images were selected for data augmentation and application of Gaussian noise. 10 minutes -3candle resulted in 79.72% accuracy. The accuracy of the images with 0.00025 modified value and 100% changed candles was 79.92%. Applying Gaussian noise helped the accuracy to be 80.98%. According to the outcomes of Case 2, 60minutes candle charts could predict patterns of tomorrow by 82.60%. To sum up, this study is expected to contribute to further studies on the prediction of stock price patterns using images. This research provides a possible method for data augmentation of stock data.

• Fashion & Textile Research Journal
• /
• v.1 no.3
• /
• pp.252-258
• /
• 1999

The aims of this study is to classify the lower somatotype of adult women and appearance analysis on the shape of flare skirts by using the image processing. Also We have made skirts in order to analyze the various types of appearance of flare skirts by using the image processing. The subjects for our wear test lower somatotype, who were controlled in their waist, abdomen and hip shapes. The flare skirts used for wear test were 112 types(combinated 14 fabric type and 8 lower somatotype). The effect of lower somatotype on the shape of flare skirts was determined by the horizontally hem line section shape and the silhouette of flare skirts with image processing. The data were analyzed by using analysis of variance and Turkey, Duncan multiple range test. The results obtained are summarized as follows: It is shown that the fabric weight elongation differs in fabrics properties, in direction of textures. The shape horizontal section of flare skirt hem line has differed with the number of nodes, wave-height of nodes and breadth of silhouette by fabrics properties and lower somatotype. It is noticed that the breadth of flare skirts by the silhouette has high correlation with the drape ability of fabrics and lower somatotype. Results for our flare skirts show changes in amplitude and mean with fabrics, somatotype. Therefore we can say that gray-level histograms are correlated with changes in appearance, differences in drape spacing and related fabric properties and their somatotype.

• Journal of Korean Academy of Fundamentals of Nursing
• /
• v.6 no.3
• /
• pp.414-433
• /
• 1999

The purpose of this study is to identify the needs which are perceived by one of the familiy members who have head injury parients by traffic accidents in the intensive care units. Subjects were 70 families members of head injury patients admitted to 2 general hospitals NICU located Pusan city, 1 general hospital NICU located Ulsan city from December 1. 1998 to February 28. 1999. For this study, 70 family members were interviewed with aid of a Likert scale which was developed by researcher for this study. For development of the survey instrument, literature review and open questionnaire technique with family members and the nurses working in NICU. The 50needs-items were classified into 7 groups according to the homogeneity of the items with the support of literature review. For the content validity, the instrument was reviewed by 1 nursing professor and the internal reliability of this instrument was Cronbach alpha=0.94 which is highly accepted. Data was analyzed by a SPSS computer program. Data analysis included frequency. percentage, mean, standard variance and t-test or ANOVA. The results were as followings : 1. The general characteristics of head injury patients shows that the male was 74.3%, the female was 25.7% and age distribution shows that the fifty-fifty nine years was 30%, the highest. Of religion the buddhism was the most, The diagnosis distribution shows that epidural hematoma was 32.9% and subdural hematoma was 24.3%. The mentalility distribution shows that semicoma was 31.4% and stupor was 31.4%. Hemiplegia was 42.9% 2. The general charaterisrics of the family needs of head injury patients shows that thirty-thirty nine years was 31.4%, the highest. sex distribution shows that the male was 20%, the female was 80%. Of religion the buddism was the most. 3. The family needs of head injury patients was $3.03{\pm}0.42$, needs for the information of a patient's condition was $3.65{\pm}0.48$, the highest. And needs for the information of care and treatment was $3.48{\pm}0.48$, needs to be supplied with comfortable facilities for family was $3.04{\pm}0.66$, needs to be participate in a patient's care was $2.90{\pm}0.55$, needs to be informed about the available resources was $2.83{\pm}0.59$, needs to be supported emotionally for family was $2.79{\pm}0.55$, needs for religious assistance was $2.51{\pm}0.85$. 4. Examining the family needs of head injury patients according to patient's characteristics, mentality, plegia degree and operation were shown be variables to make an effect on the needs for the head injury patients family. At all, according to severity of head. injury, the family needs of head injury patients was high. 5. Examining the family needs of head injury patients according to their general characteristics, we could know that religion, job. income were shown to be variables to make an effect on the family needs. Through the examination it can be seen that the characteristics of head injury patients and the family needs of head injury patients. In conclusion, the family needs of head injury patients was almost same the family needs of ICU patients. Therefore we must involve the family's care of head injury patients and we must provide exact and repeated explanation, education and support the family of head injury patients. As this study was resulted in selecting the families admitted to NICU of some general hospital, we couldn't stretch the result in our favor. Therefore, continuous studies are suggested.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.22 no.3
• /
• pp.21-36
• /
• 2019

Dead fuel moisture content is a key variable in fire danger rating as it affects fire ignition and behavior. This study evaluates simple regression models estimating the moisture content of standardized 10-h fuel stick (10-h FMC) at three sites with different characteristics(urban and outside/inside the forest). Equilibrium moisture content (EMC) was used as an independent variable, and in-situ measured 10-h FMC was used as a dependent variable and validation data. 10-h FMC spatial distribution maps were created for dates with the most frequent fire occurrence during 2013-2018. Also, 10-h FMC values of the dates were analyzed to investigate under which 10-h FMC condition forest fire is likely to occur. As the results, fitted equations could explain considerable part of the variance in 10-h FMC (62~78%). Compared to the validation data, the models performed well with R2 ranged from 0.53 to 0.68, root mean squared error (RMSE) ranged from 2.52% to 3.43%, and bias ranged from -0.41% to 1.10%. When the 10-h FMC model fitted for one site was applied to the other sites, $R^2$ was maintained as the same while RMSE and bias increased up to 5.13% and 3.68%, respectively. The major deficiency of the 10-h FMC model was that it poorly caught the difference in the drying process after rainfall between 10-h FMC and EMC. From the analysis of 10-h FMC during the dates fire occurred, more than 70% of the fires occurred under a 10-h FMC condition of less than 10.5%. Overall, the present study suggested a simple model estimating 10-h FMC with acceptable performance. Applying the 10-h FMC model to the automatic mountain weather observation system was successfully tested to produce a national-scale 10-h FMC spatial distribution map. This data will be fundamental information for forest fire research, and will support the policy maker.

• Journal of Korean Society of Forest Science
• /
• v.89 no.5
• /
• pp.645-654
• /
• 2000

The aim of this study was to described the genetic structure of Eleutherococcus senticosus in Korea. We investigated 10 patches, which are eight Korean patches and two foreign patches come from Russia and China growing at Korean habitat, using ISSR(inter-simple sequence repeats) markers. In ISSR PCR, the overall percentage of polymorphic ISSR amplicons was 76% and the mean number of amplicons per ISSR primer was 11.5, which were higher than the RAPD results for the some cultivars collected in Korea(Kim et al., 1998) ; 57% and 5.7, respectively. So ISSR markers provide more powerful tool than RAPD markers for the investigation of genetic variation in E. senticosus. There are relatively high genetic variation among patches as 62.8%, but low variation within eight Korean patches. Such pattern of genetic variation, which is not ordinary in other tree species, may be result from the narrow and limited habitats and the asexual reproduction of this species at the natural stands in Korea. Although the small sample size in this study seemed to be resulted in the high genetic variation among patches, the overall genetic interpretation of this study might not be much affected on the basis of the characteristics of the distribution and the reproduction system of E. senticosus. Analysis of genetic distance between all pairs of the patches did not reveal any trends with regard to geographic distance, which was confirmed by the results obtained from AMOVA(analysis of molecular variance) and PCA(principal component analysis). These results suggest that, in addition to the preservation of the natural stands, the conservation of larger number of patches with small number of individuals per patch is more effective for the ex situ conservation and for maintaining the genetic diversity of E. senticosus in Korea than smaller patches with large number of individuals.

• Progress in Medical Physics
• /
• v.25 no.1
• /
• pp.37-45
• /
• 2014

The accuracy and uniformity of CT numbers are the main causes of radiation dose calculation error. Especially, for the dose calculation based on kV-Cone Beam Computed Tomography (CBCT) image, the scatter affecting the CT number is known to be quite different by the object sizes, densities, exposure conditions, and so on. In this study, the scatter impact on the CBCT based dose calculation was evaluated to provide the optimal condition minimizing the error. The CBCT images was acquired under three scatter conditions ("Under-scatter", "Over-scatter", and "Full-scatter") by adjusting amount of scatter materials around a electron density phantom (CIRS062, Tissue Simulation Technology, Norfolk, VA, USA). The CT number uniformities of CBCT images for water-equivalent materials of the phantom were assessed, and the location dependency, either "inner" or "outer" parts of the phantom, was also evaluated. The electron density correction curves were derived from CBCT images of the electron density phantom in each scatter condition. The electron density correction curves were applied to calculate the CBCT based doses, which were compared with the dose based on Fan Beam Computed Tomography (FBCT). Also, 5 prostate IMRT cases were enrolled to assess the accuracy of dose based on CBCT images using gamma index analysis and relative dose differences. As the CT number histogram of phantom CBCT images for water equivalent materials was fitted with a gaussian function, the FHWM (146 HU) for "Full-scatter" condition was the smallest among the FHWM for the three conditions (685 HU for "under scatter" and 264 HU for "over scatter"). Also, the variance of CT numbers was the smallest for the same ingredients located in the center and periphery of the phantom in the "Full-scatter" condition. The dose distributions calculated with FBCT and CBCT images compared in a gamma index evaluation of 1%/3 mm criteria and in the dose difference. With the electron density correction acquired in the same scatter condition, the CBCT based dose calculations tended to be the most accurate. In 5 prostate cases in which the mean equivalent diameter was 27.2 cm, the averaged gamma pass rate was 98% and the dose difference confirmed to be less than 2% (average 0.2%, ranged from -1.3% to 1.6%) with the electron density correction of the "Full-scatter" condition. The accuracy of CBCT based dose calculation could be confirmed that closely related to the CT number uniformity and to the similarity of the scatter conditions for the electron density correction curve and CBCT image. In pelvic cases, the most accurate dose calculation was achievable in the application of the electron density curves of the "Full-scatter" condition.

• Journal of Environmental Impact Assessment
• /
• v.27 no.4
• /
• pp.353-366
• /
• 2018

The purpose of this study was assessment the quality of water by using the statistical analysis technique of the Water flow and water quality from January 2012 to December 2016 at the unit basin for total pollutant load management system (TPLMS) in the Imjin River. Water flow and water quality were monitored at an average of 8 day intervals, 11 parameters were used for correlation analysis, principal component analysis (PCA), factor analysis (FA), and cluster analysis (CA). The Hierarchical CA was classified into three according to the change of space, such as natural rivers, urban rivers, point with large influence of point pollution source, it was found that the type of contamination source the similarity of water quality affected the classification of cluster. Using one-way analysis of variance (ANOVA) and post-hoc Analysis, there were statistically significant differences between mean values among the clusters. Correlation analysis showed the correlation coefficient between $COD_{Mn}$ and TOC was 0.951 (p<0.01) and the correlation was statistically significantly higher. According to the result PCA and FA, 3 principal components can explaining 72% of the total variations in water quality characteristics and main factor was EC, $BOD_5$, $COD_{Mn}$, TN, TP and TOC indirect indicators of organic matter and nutrients were influenced. This study presented the regression equation obtained by applying the factor scores to the multiple linear regression analysis and concluded that the management Indirect indicators of organic matter and nutrients is important for water quality management in the Imjin River basin.

• Journal of agricultural medicine and community health
• /
• v.41 no.1
• /
• pp.13-26
• /
• 2016

Purpose: This study was performed to identify the association between family support, activities of daily living (ADL) and depression among hospitalized older patients with chronic diseases. Methods: This study subjects were 100 elderly patients with chronic diseases including chronic respiratory diseases, diabetes mellitus and et al. in a general hospital. The collected data were patient characteristics, family support, ADL, and depression by structured questionnaire and medical chart review. The used statistical analyses were t-test, analysis of variance, Pearson's correlational analysis and multiple regression analysis. Results: The mean scores of family support, ADL and depression were $49.95{\pm}8.68$, $8.65{\pm}2.65$, $6.66{\pm}3.78$, respectively. The prevalence rate of depression was 64.0%. In simple analysis, the statistically significant associated factors with depression were age, spouse, economic status, social activity, subjective health status, and number of pain. Depression had statistically a significant positive correlation with ADL and a negative correlation with family support. The final result of hierarchial multiple regression analysis (Model 3), the factors related to depression were family support (b=-.135, p<.001), subjective health status (b=2.510, p=.001). Conclusions: It is necessary to develop and apply the program for controlling the depression of elderly patients with health education, reinforcement of supportive systems in hospital. And, further multidisciplinary studies should be done.

• Journal of Korean Ophthalmic Optics Society
• /
• v.18 no.3
• /
• pp.231-239
• /
• 2013

Purpose: To design and fabricate the auto pattern maker for the development. Methods: we got the necessary data, needed in design, by using CAD. Based on the these data, we fabricated the trial product for the development of the auto pattern maker. Results: The auto pattern maker were composed with combinations of many elements; pattern making assembly, control panel, frame attachment and prober unit. The pattern making assembly was comprised of the cutter, the pattern holder, pattern remover and silence cover which could minimize the sound during the cutting process. The control panel was designed to be connected and operated with the main printed circuit board. The prober could get the eye shape data by scanning of 1.8 degrees around the groove of the frame through the encoding data according to the address. After starting, scanning was carried out in two passes, i.e. one right-handed and one left-handed. Communication connector could send the eye shape data from auto pattern maker to outer system with the RS232C transmission system. By using the one-way analysis of variance, we got the error rate of cut pattern size for ${\Phi}22mm$, ${\Phi}55mm$ and ${\Phi}62mm$. Because F-value was 0.510 and p-value was 0.601, no statistically significant differences were found. Also, the mean cutting error of the auto pattern maker was 0.0274 mm. Conclusions: we could succeed in making the trial product by applying it to the development of the auto pattern maker. The role of this auto pattern maker is to find a exact required size of lens to fit the frame by measuring the frame. The acquired data are transferred to outer system for grinding and finishing with patternless process. Also, the trial product can produce pattern to fit the frame. Therefore, it was confidently expected that the optometrists could handily produce pattern to fit the frame with this trial product and dispense the ophthalmic lens because of its efficiency and convenience compared to the past.

• Journal of Dental Rehabilitation and Applied Science
• /
• v.24 no.3
• /
• pp.231-242
• /
• 2008

Purpose: The purpose of this study was to investigate the bond strength of the core-veneer interface in all ceramic systems. Material and Methods: The all ceramic systems tested with their respective veneer were IPS Empress 2 with IPS Eris, IPS e.max Press with IPS e.max Ceram and IPS-e.max ZirCAD with IPS e.max Ceram. Cores (N=36, N=12/group, diameter: 10mm, thickness: 3mm) were fabricated according to the manufacturer's instruction and cleaned with ultrasonic cleaner. The veneer(diameter: 3mm, thickness: 2mm) were condensed in stainless steel mold and fired on to the core materials. After firing, they were again ultrasonically cleaned and embedded in acrylic resin. The specimens were stored in distilled water at $37^{\circ}C$ for 1 week. The specimens were placed in a mounting jig and subjected to shear force in a universal testing machine(Z020, Zwick, Germany). Load was applied at close to the core-veneer interface as possible with crosshead speed of 1.00mm/min until failure. Average shear bond strengths(MPa) were analyzed with a one-way analysis of variance and the Tukey test(${\alpha}=.05$). The failed specimens were examinated by scanning electron microscopy(JSM-6360, JEOL, Japan). The pattern of failure was classified as cohesive in core, cohesive in veneer, mixed or adhesive. Results: The mean shear bond strength($MPa{\pm}SD$) were IPS e.max Press $32.85{\pm}6.75MPa$, IPS Empress 2 $29.30{\pm}6.51MPa$, IPS e.max ZirCAD $28.10{\pm}4.28MPa$. IPS Empress 2, IPS e.max Press, IPS e.max ZirCAD were not significantly different from each others. Scanning electron microscopy examination revealed that adhesive failure did not occur in any all ceramic systems. IPS Empress 2 and IPS e.max Press exhibited cohesive failure in both the core and the veneer. IPS e.max ZirCAD exhibited cohesive failure in veneer and mixed failure.