• Journal of Intelligence and Information Systems
• /
• v.24 no.1
• /
• pp.167-181
• /
• 2018

Over the past decade, deep learning has been in spotlight among various machine learning algorithms. In particular, CNN(Convolutional Neural Network), which is known as the effective solution for recognizing and classifying images or voices, has been popularly applied to classification and prediction problems. In this study, we investigate the way to apply CNN in business problem solving. Specifically, this study propose to apply CNN to stock market prediction, one of the most challenging tasks in the machine learning research. As mentioned, CNN has strength in interpreting images. Thus, the model proposed in this study adopts CNN as the binary classifier that predicts stock market direction (upward or downward) by using time series graphs as its inputs. That is, our proposal is to build a machine learning algorithm that mimics an experts called 'technical analysts' who examine the graph of past price movement, and predict future financial price movements. Our proposed model named 'CNN-FG(Convolutional Neural Network using Fluctuation Graph)' consists of five steps. In the first step, it divides the dataset into the intervals of 5 days. And then, it creates time series graphs for the divided dataset in step 2. The size of the image in which the graph is drawn is $40(pixels){\times}40(pixels)$, and the graph of each independent variable was drawn using different colors. In step 3, the model converts the images into the matrices. Each image is converted into the combination of three matrices in order to express the value of the color using R(red), G(green), and B(blue) scale. In the next step, it splits the dataset of the graph images into training and validation datasets. We used 80% of the total dataset as the training dataset, and the remaining 20% as the validation dataset. And then, CNN classifiers are trained using the images of training dataset in the final step. Regarding the parameters of CNN-FG, we adopted two convolution filters ($5{\times}5{\times}6$ and $5{\times}5{\times}9$) in the convolution layer. In the pooling layer, $2{\times}2$ max pooling filter was used. The numbers of the nodes in two hidden layers were set to, respectively, 900 and 32, and the number of the nodes in the output layer was set to 2(one is for the prediction of upward trend, and the other one is for downward trend). Activation functions for the convolution layer and the hidden layer were set to ReLU(Rectified Linear Unit), and one for the output layer set to Softmax function. To validate our model - CNN-FG, we applied it to the prediction of KOSPI200 for 2,026 days in eight years (from 2009 to 2016). To match the proportions of the two groups in the independent variable (i.e. tomorrow's stock market movement), we selected 1,950 samples by applying random sampling. Finally, we built the training dataset using 80% of the total dataset (1,560 samples), and the validation dataset using 20% (390 samples). The dependent variables of the experimental dataset included twelve technical indicators popularly been used in the previous studies. They include Stochastic %K, Stochastic %D, Momentum, ROC(rate of change), LW %R(Larry William's %R), A/D oscillator(accumulation/distribution oscillator), OSCP(price oscillator), CCI(commodity channel index), and so on. To confirm the superiority of CNN-FG, we compared its prediction accuracy with the ones of other classification models. Experimental results showed that CNN-FG outperforms LOGIT(logistic regression), ANN(artificial neural network), and SVM(support vector machine) with the statistical significance. These empirical results imply that converting time series business data into graphs and building CNN-based classification models using these graphs can be effective from the perspective of prediction accuracy. Thus, this paper sheds a light on how to apply deep learning techniques to the domain of business problem solving.

• Journal of Intelligence and Information Systems
• /
• v.27 no.3
• /
• pp.139-156
• /
• 2021

The biggest reason for using a deep learning model in image classification is that it is possible to consider the relationship between each region by extracting each region's features from the overall information of the image. However, the CNN model may not be suitable for emotional image data without the image's regional features. To solve the difficulty of classifying emotion images, many researchers each year propose a CNN-based architecture suitable for emotion images. Studies on the relationship between color and human emotion were also conducted, and results were derived that different emotions are induced according to color. In studies using deep learning, there have been studies that apply color information to image subtraction classification. The case where the image's color information is additionally used than the case where the classification model is trained with only the image improves the accuracy of classifying image emotions. This study proposes two ways to increase the accuracy by incorporating the result value after the model classifies an image's emotion. Both methods improve accuracy by modifying the result value based on statistics using the color of the picture. When performing the test by finding the two-color combinations most distributed for all training data, the two-color combinations most distributed for each test data image were found. The result values were corrected according to the color combination distribution. This method weights the result value obtained after the model classifies an image's emotion by creating an expression based on the log function and the exponential function. Emotion6, classified into six emotions, and Artphoto classified into eight categories were used for the image data. Densenet169, Mnasnet, Resnet101, Resnet152, and Vgg19 architectures were used for the CNN model, and the performance evaluation was compared before and after applying the two-stage learning to the CNN model. Inspired by color psychology, which deals with the relationship between colors and emotions, when creating a model that classifies an image's sentiment, we studied how to improve accuracy by modifying the result values based on color. Sixteen colors were used: red, orange, yellow, green, blue, indigo, purple, turquoise, pink, magenta, brown, gray, silver, gold, white, and black. It has meaning. Using Scikit-learn's Clustering, the seven colors that are primarily distributed in the image are checked. Then, the RGB coordinate values of the colors from the image are compared with the RGB coordinate values of the 16 colors presented in the above data. That is, it was converted to the closest color. Suppose three or more color combinations are selected. In that case, too many color combinations occur, resulting in a problem in which the distribution is scattered, so a situation fewer influences the result value. Therefore, to solve this problem, two-color combinations were found and weighted to the model. Before training, the most distributed color combinations were found for all training data images. The distribution of color combinations for each class was stored in a Python dictionary format to be used during testing. During the test, the two-color combinations that are most distributed for each test data image are found. After that, we checked how the color combinations were distributed in the training data and corrected the result. We devised several equations to weight the result value from the model based on the extracted color as described above. The data set was randomly divided by 80:20, and the model was verified using 20% of the data as a test set. After splitting the remaining 80% of the data into five divisions to perform 5-fold cross-validation, the model was trained five times using different verification datasets. Finally, the performance was checked using the test dataset that was previously separated. Adam was used as the activation function, and the learning rate was set to 0.01. The training was performed as much as 20 epochs, and if the validation loss value did not decrease during five epochs of learning, the experiment was stopped. Early tapping was set to load the model with the best validation loss value. The classification accuracy was better when the extracted information using color properties was used together than the case using only the CNN architecture.

• Korean Journal of Soil Science and Fertilizer
• /
• v.10 no.3
• /
• pp.103-134
• /
• 1977

The physiological and biochemical role of potassium for upland crops according to recent research reports and the nutritional status of potassium in Korea were reviewed. Since physical and chemical characteristics of potassium ion are different from those of sodium, potassium can not completely be replaced by sodium and replacement must be limited to minimum possible functional area. Specific roles of potassium seem to keep fine structure of biological membranes such as thylacoid membrane of chloroplast in the most efficient form and to be allosteric effector and conformation controller of various enzymes principally in carbohydrate and protein metabolism. Potassium is essential to improve the efficiency of phoro- and oxidative- phosphorylation and involve deeply in all energy required metabolisms especially synthesis of organic matter and their translocation. Potassium has many important, physiological functions such as maintenance of osmotic pressure and optimum hydration of cell colloids, consequently uptake and translocation of water resulting in higher water use efficiency and of better subcellular environment for various physiological and biochemical activities. Potassium affects uptake and translocation of mineral nutrients and quality of products. potassium itself in products may become a quality criteria due to potassium essentiality for human beings. Potassium uptake is greatly decreased by low temperature and controlled by unknown feed back mechanism of potassium in plants. Thus the luxury absorption should be reconsidered. Total potassium content of upland soil in Korea is about 3% but the exchangeable one is about 0.3 me/100g soil. All upland crops require much potassium probably due to freezing and cold weather and also due to wet damage and drought caused by uneven rainfall pattern. In barley, potassium should be high at just before freezing and just after thawing and move into grain from heading for higher yield. Use efficiency of potassium was 27% for barley and 58% in old uplands, 46% in newly opened hilly lands for soybean. Soybean plant showed potassium deficiency symptom in various fields especially in newly opened hilly lands. Potassium criteria for normal growth appear 2% $K_2O$ and 1.0 K/(Ca+Mg) (content ratio) at flower bud initiation stage for soybean. Potassium requirement in plant was high in carrot, egg plant, chinese cabbage, red pepper, raddish and tomato. Potassium content in leaves was significantly correlated with yield in chinese cabbage. Sweet potato. greatly absorbed potassium subsequently affected potassium nutrition of the following crop. In the case of potassium deficiency, root showed the greatest difference in potassium content from that of normal indicating that deficiency damages root first. Potatoes and corn showed much higher potassium content in comparison with calcium and magnesium. Forage crops from ranges showed relatively high potassium content which was significantly and positively correlated with nitrogen, phosphorus and calcium content. Percentage of orchards (apple, pear, peach, grape, and orange) insufficient in potassium ranged from 16 to 25. The leaves and soils from the good apple and pear orchards showed higher potassium content than those from the poor ones. Critical ratio of $K_2O/(CaO+MgO)$ in mulberry leaves to escape from winter death of branch tip was 0.95. In the multiple croping system, exchangeable potassium in soils after one crop was affected by the previous crops and potassium uptake seemed to be related with soil organic matter providing soil moisture and aeration. Thus, the long term and quantitative investigation of various forms of potassium including total one are needed in relation to soil, weather and croping system. Potassium uptake and efficiency may be increased by topdressing, deep placement, slow-releasing or granular fertilizer application with the consideration of rainfall pattern. In all researches for nutritional explanation including potassium of crop yield reasonable and practicable nutritional indices will most easily be obtained through multifactor analysis.

• The Southeast Asian review
• /
• v.27 no.2
• /
• pp.37-76
• /
• 2017

$Nguy{\tilde{\hat{e}}}n$ Cư Trinh has two faces in the history of territorial expansion of Vietnam into the Mekong delta. One is his heroic contribution to the $Nguy{\tilde{\hat{e}}}n$ family gaining control over the large part of the Mekong delta. The other is his role to make the eyes of readers of Vietnamese history be fixed only to the present territory of Vietnam. To the readers, $Nguy{\tilde{\hat{e}}}n$ Cư Trinh's achievement of territorial expansion was the final stage of the nam $ti{\acute{\hat{e}}n$ of Vietnam. In fact, however, his achievement was partial. This study pays attention to the King $V{\tilde{o}}$ instead of $Nguy{\tilde{\hat{e}}}n$ Cư Trinh in the history of the territorial expansion in the Mekong delta. King's goal was more ambitious. And the ambition was propelled by his dream to build a new world, and its order, in which his new capital, $Ph{\acute{u}}$ $Xu{\hat{a}}n$ was to be the center with his status as an emperor. To improve my assertion, three elements were examined in this article. First is the nature of $V{\tilde{o}}$ Vương's new kingship. Second is the preparation and the background of the military operation in the Mekong Delta. The nature of the new territory is the third element of the discussion. In 1744, six years after this ascending to the throne, $V{\tilde{o}}$ Vương declared he was a king. Author points out this event as the departure of the southern kingdom from the traditional dynasties based on the Red River delta. Besides, the government system, northern custom and way of dressings were abandoned and new southern modes were adopted. $V{\tilde{o}}$ Vương had enough tributary kingdoms such as Cambodia, Champa, Thủy $X{\tilde{a}}$, Hoả $X{\tilde{a}}$, Vạn Tượng, and Nam Chưởng. Compared with the $L{\hat{e}}$ empire, the number of the tributary kingdoms was higher and the number was equivalent to that of the Đại Nam empire of the 19th century. In reality, author claims, the King $V{\tilde{o}}^{\prime}s$ real intention was to become an emperor. Though he failed in using the title of emperor, he distinguished himself by claiming himself as the Heaven King, $Thi{\hat{e}}n$ Vương. Cambodian king's attack on the thousands of Cham ethnics in Cambodian territory was an enough reason to the King $V{\tilde{o}}^{\prime}s$ military intervention. He considered these Cham men and women as his amicable subjects, and he saw them a branch of the Cham communities in his realm. He declared war against Cambodia in 1750. At the same time he sent a lengthy letter to the Siamese king claiming that the Cambodia was his exclusive tributary kingdom. Before he launched a fatal strike on the Mekong delta which had been the southern part of Cambodia, $V{\tilde{o}}$ Vương renovated his capital $Ph{\acute{u}}$ $Xu{\hat{a}}n$ to the level of the new center of power equivalent to that of empire for his sake. Inflation, famine, economic distortion were also the features of this time. But this study pays attention more to the active policy of the King $V{\tilde{o}}$ as an empire builder than to the economic situation that has been told as the main reason for King $V{\tilde{o}}^{\prime}s$ annexation of the large part of the Mekong delta. From the year of 1754, by the initiative of $Nguy{\tilde{\hat{e}}}n$ Cư Trinh, almost whole region of the Mekong delta within the current border line was incorporated into the territory of $V{\tilde{o}}$ Vương within three years, though the intention of the king was to extend his land to the right side of the Mekong Basin beyond the current border such as Kampong Cham, Prey Vieng, and Svai Rieng. The main reason was $V{\tilde{o}}$ Vương's need to expand his territory to be matched with that of his potential empire with the large number of the tributary kingdoms. King $V{\tilde{o}}^{\prime}s$ strategy was the variation of 'silkworm nibbling' and 'to strike barbarians by barbarians.' He ate the land of Lower Cambodia, the region of the Mekong delta step by step as silkworm nibbles mulberry leave(general meaning of $t{\acute{a}}m$ thực), but his final goal was to eat all(another meaning of $t{\acute{a}}m$ thực) the part of the Mekong delta including the three provinces of Cambodia mentioned above. He used Cham to strike Cambodian in the process of getting land from Long An area to $Ch{\hat{a}}u$ Đốc. This is a faithful application of the Dĩ Man $C{\hat{o}}ng$ Man (to strike barbarians by barbarians). In addition he used Chinese refugees led by the Mạc family or their quasi kingdom to gain land in the region of $H{\grave{a}}$ $Ti{\hat{e}}n$ and its environs from the hand of Cambodian king. This is another application of Dĩ Man $C{\hat{o}}ng$ Man. In sum, author claims a new way of looking at the origin of the imperial world order which emerged during the first half of the 19th century. It was not the result of the long history of Đại Việt empires based on the Red River delta, but the succession of the King $V{\tilde{o}}^{\prime}s$ new world based on $Ph{\acute{u}}$ $Xu{\hat{a}}n$. The same ways of Dĩ Man $C{\hat{o}}ng$ Man and $T{\acute{a}}m$ Thực Chi $K{\acute{\hat{e}}}$ were still used by $V{\tilde{o}}^{\prime}s$ descendents. His grandson Gia Long used man such as Thai, Khmer, Lao, Chinese, and European to win another man the '$T{\hat{a}}y$ Sơn bandits' that included many of Chinese pirates, Cham, and other mountain peoples. His great grand son Minh Mạng constructed a splendid empire. At the same time, however, Minh Mạng kept expanding the size of his empire by eating all the part of Cambodia and Cham territories.

• Korean Journal of Heritage: History & Science
• /
• v.45 no.2
• /
• pp.20-37
• /
• 2012

After withdrawal of military troops of Chinese Tang dynasty in the 18th year of King Moon-moo's reign(678), the Silla Kingdom had actually unified the Korean peninsula and had divided the territory into 9 states benchmarking the China's local administrations adjustment system. He had established local administrative units by deploying secondary capitals, counties and prefectures in the nine states. The so-called "9 Provinces and 5 Secondary capitals" are what constitutes the local administrations system. The provinces can be compared to current provinces of the Republic of Korea(hereinafter Korea), and secondary capitals to megalopolises. According to a chapter of the Samkuksaki(三?史記) which had recorded the achievements of king Kyoungdeok in December in his 16th year on the throne(757), the local administrative units had amounted to 5 secondary capitals, 117 counties and 293 prefectures. There are still lots of ambiguous points since there have never been any consultation on locations of provinces and secondary capitals' castles, and on structures of cities because the researches for local cities inside the 9 Provinces and 5 Secondary capitals in the Unified Silla Kingdom has been conducted centering on the historic literatures only. The research for restoring structures of cities seen from an archeological perspective are limited to the studies of Taewoo Park("A study on the local cities in the Unified Kingdom Age" 1987) and that of the author("A study on the restoration of planned cities for the Unified Silla Kingdom in terms of the structures and realities of the castles in the 9 Provinces and 5 Secondary capitals" 2009). The Gangneung city of Gangwon province was originally called Haseoryang(河西良) of the Gogureo Kingdom as an ancient nation of Ye(濊). According to "Samkuksaki", it had evolved from Haseoju(河西州) to a secondary capitals in the 8th year of King Seonduk(639). Afterwards, it had been renamed as Myeongju(溟洲) in the 16th year of King Kyoungduk(757), and then several other names were given to it after Goryo dynasty. Taewoo Park claims that it is being defined as a sanctuary remaining in Myoungjudong because of the vestige of bare castle, and this cannot be ascertained due to the on-going urbanization processes. Also, the Kwandong university authority is suggesting an opinion of regarding Myeongju mountain castle located 3 Kms southwest of the center of Gangwon city as commanding post for the pertinent state. The author has restored the pertinent area into a city composed of villages within a lattice framework like Silla Keumkyoung and many other cities. The structure is depicted next. The downtown of Gangneung is situated on a flat terrain at the west bank of Namdaecheon stream flowing southwest to northeast along the inner area of the city. Though there isn't any hill comparatively higher than others in the vicinity, hills are continuously linked east to west along the northern area of the downtown, and the maximum width of flat terrain is about 1 Km and is not so large. Currently, urbanization is being proceeded into the inner portion of Gangneung city, the lands in all directions from the hub of Gangneung station have been readjusted, and thus previous land-zoning program is almost nullified. However, referring to the topographic chart drawn at the time of Japanese colonial rule, it can be validated that land-zoning program to accord the lattice framework with the length of its one side equaling to 190m leaves its vestige about 0.8Km northwest to southeast and about 1.7Km northeast to southwest of the vicinity of Okcheondong, Imdangdong, Geumhakdong, Myeongjudong, and etcetera which comprize the hub of the downtown. The land-zoning vestige within the lattice framework, compared to other cases related with the '9 states and 5 secondary capitals', is very much likely to be that of the Unified Silla Kingdom. That the length of a side of a lattice framework is 190m as opposed to that of Silla Geumkyoung and other cities with their 140m or 160m long sides is a single survey item in the future. The baseline direction for zoning the lands is tilting approximately 37.5 degrees west of northwest to southeast axis in accordance with the topographic features. It seems that this phenomenon takes place because of the direction of Namdaecheon and the geographic constraints of the hills in the north. Reviewing minimally, a rectangular size of zoned land by 4 Pangs(坊) on the northwest to southeast side multiplied by 7 Pangs(坊) on the northeast to southwest side had been restored within a lattice framework. Otherwise, considering the extent of expansion of the existing zoned lands in the lattice framework and one more Pang(坊) being added to each side, it is likely that the size could have been with 5 Pangs(坊) on the northwest to southeast side multiplied by 8 Pangs(坊) on the northeast to southwest side(950 M on the northwest to southeast side multiplied by 1,520m on the northeast to southwest side). The overall shape is rectangle, but land-zoning programs reminiscent of rebuilt roads(red phoenix road) like Jang-an castle(長安城) of Chinese Tang dynasty or Pyoungseong castle(平城城) in Japan is not to be validated. There are some historic items among the roof tiles and earthen wares excavated at local administrative office sites or Gangneung's town castle in Joseon dynasty inside the area assumed to be containing municipal vestiges even though archeological survey for the vestige of Myeongju has not been made yet, and these items deserve dating back to the Unified Silla Kingdom age. Also, all of the construction sites at local administrative authorities of the Joseon dynasty are showing large degrees of slant in the azimuth. This is a circumstantial evidence indicating the fact that the inherited land-zoning programs to be seen in Gangneung in terms of the lattice framework had ever existed in the past. Also, the author does not decline that Myeongju mountain castle had once been the commanding post when reviewing the roof tiles at the edge of eaves in this stronghold. The ancient municipal castles in the Korean peninsula are composed of castles on the flat terrain as well as hilly areas and the cluster of strongholds like Myounghwal, Namhan, Seohyoung mountain castles built around municipal castle of Geumkyoung based on a lattice framework program. Considering that mountain castles are spread in the vicinity of municipal vestiges in other cities other than the 9 states and 5 secondary capitals, it is estimated that Myeongju was assuming the function of commanding post incorporating cities on the flat terrain and castles on the hills.