• Journal of Intelligence and Information Systems
• /
• v.24 no.1
• /
• pp.1-23
• /
• 2018

From the 21st century, various high-quality services have come up with the growth of the internet or 'Information and Communication Technologies'. Especially, the scale of E-commerce industry in which Amazon and E-bay are standing out is exploding in a large way. As E-commerce grows, Customers could get what they want to buy easily while comparing various products because more products have been registered at online shopping malls. However, a problem has arisen with the growth of E-commerce. As too many products have been registered, it has become difficult for customers to search what they really need in the flood of products. When customers search for desired products with a generalized keyword, too many products have come out as a result. On the contrary, few products have been searched if customers type in details of products because concrete product-attributes have been registered rarely. In this situation, recognizing texts in images automatically with a machine can be a solution. Because bulk of product details are written in catalogs as image format, most of product information are not searched with text inputs in the current text-based searching system. It means if information in images can be converted to text format, customers can search products with product-details, which make them shop more conveniently. There are various existing OCR(Optical Character Recognition) programs which can recognize texts in images. But existing OCR programs are hard to be applied to catalog because they have problems in recognizing texts in certain circumstances, like texts are not big enough or fonts are not consistent. Therefore, this research suggests the way to recognize keywords in catalog with the Deep Learning algorithm which is state of the art in image-recognition area from 2010s. Single Shot Multibox Detector(SSD), which is a credited model for object-detection performance, can be used with structures re-designed to take into account the difference of text from object. But there is an issue that SSD model needs a lot of labeled-train data to be trained, because of the characteristic of deep learning algorithms, that it should be trained by supervised-learning. To collect data, we can try labelling location and classification information to texts in catalog manually. But if data are collected manually, many problems would come up. Some keywords would be missed because human can make mistakes while labelling train data. And it becomes too time-consuming to collect train data considering the scale of data needed or costly if a lot of workers are hired to shorten the time. Furthermore, if some specific keywords are needed to be trained, searching images that have the words would be difficult, as well. To solve the data issue, this research developed a program which create train data automatically. This program can make images which have various keywords and pictures like catalog and save location-information of keywords at the same time. With this program, not only data can be collected efficiently, but also the performance of SSD model becomes better. The SSD model recorded 81.99% of recognition rate with 20,000 data created by the program. Moreover, this research had an efficiency test of SSD model according to data differences to analyze what feature of data exert influence upon the performance of recognizing texts in images. As a result, it is figured out that the number of labeled keywords, the addition of overlapped keyword label, the existence of keywords that is not labeled, the spaces among keywords and the differences of background images are related to the performance of SSD model. This test can lead performance improvement of SSD model or other text-recognizing machine based on deep learning algorithm with high-quality data. SSD model which is re-designed to recognize texts in images and the program developed for creating train data are expected to contribute to improvement of searching system in E-commerce. Suppliers can put less time to register keywords for products and customers can search products with product-details which is written on the catalog.

• Journal of Global Scholars of Marketing Science
• /
• v.20 no.3
• /
• pp.262-268
• /
• 2010

Virtual communities (VCs) have developed rapidly, with more and more people participating in them to exchange information and opinions. A virtual community is a group of people who may or may not meet one another face to face, and who exchange words and ideas through the mediation of computer bulletin boards and networks. A business-to-consumer virtual community (B2CVC) is a commercial group that creates a trustworthy environment intended to motivate consumers to be more willing to buy from an online store. B2CVCs create a social atmosphere through information contribution such as recommendations, reviews, and ratings of buyers and sellers. Although the importance of B2CVCs has been recognized, few studies have been conducted to examine members' word-of-mouth behavior within these communities. This study proposes a model of involvement, statistics, trust, "stickiness," and word-of-mouth in a B2CVC and explores the relationships among these elements based on empirical data. The objectives are threefold: (i) to empirically test a B2CVC model that integrates measures of beliefs, attitudes, and behaviors; (ii) to better understand the nature of these relationships, specifically through word-of-mouth as a measure of revenue generation; and (iii) to better understand the role of stickiness of B2CVC in CRM marketing. The model incorporates three key elements concerning community members: (i) their beliefs, measured in terms of their involvement assessment; (ii) their attitudes, measured in terms of their satisfaction and trust; and, (iii) their behavior, measured in terms of site stickiness and their word-of-mouth. Involvement is considered the motivation for consumers to participate in a virtual community. For B2CVC members, information searching and posting have been proposed as the main purpose for their involvement. Satisfaction has been reviewed as an important indicator of a member's overall community evaluation, and conceptualized by different levels of member interactions with their VC. The formation and expansion of a VC depends on the willingness of members to share information and services. Researchers have found that trust is a core component facilitating the anonymous interaction in VCs and e-commerce, and therefore trust-building in VCs has been a common research topic. It is clear that the success of a B2CVC depends on the stickiness of its members to enhance purchasing potential. Opinions communicated and information exchanged between members may represent a type of written word-of-mouth. Therefore, word-of-mouth is one of the primary factors driving the diffusion of B2CVCs across the Internet. Figure 1 presents the research model and hypotheses. The model was tested through the implementation of an online survey of CTrip Travel VC members. A total of 243 collected questionnaires was reduced to 204 usable questionnaires through an empirical process of data cleaning. The study's hypotheses examined the extent to which involvement, satisfaction, and trust influence B2CVC stickiness and members' word-of-mouth. Structural Equation Modeling tested the hypotheses in the analysis, and the structural model fit indices were within accepted thresholds: ${\chi}^2^$/df was 2.76, NFI was .904, IFI was .931, CFI was .930, and RMSEA was .017. Results indicated that involvement has a significant influence on satisfaction (p<0.001, ${\beta}$=0.809). The proportion of variance in satisfaction explained by members' involvement was over half (adjusted $R^2$=0.654), reflecting a strong association. The effect of involvement on trust was also statistically significant (p<0.001, ${\beta}$=0.751), with 57 percent of the variance in trust explained by involvement (adjusted $R^2$=0.563). When the construct "stickiness" was treated as a dependent variable, the proportion of variance explained by the variables of trust and satisfaction was relatively low (adjusted $R^2$=0.331). Satisfaction did have a significant influence on stickiness, with ${\beta}$=0.514. However, unexpectedly, the influence of trust was not even significant (p=0.231, t=1.197), rejecting that proposed hypothesis. The importance of stickiness in the model was more significant because of its effect on e-WOM with ${\beta}$=0.920 (p<0.001). Here, the measures of Stickiness explain over eighty of the variance in e-WOM (Adjusted $R^2$=0.846). Overall, the results of the study supported the hypothesized relationships between members' involvement in a B2CVC and their satisfaction with and trust of it. However, trust, as a traditional measure in behavioral models, has no significant influence on stickiness in the B2CVC environment. This study contributes to the growing body of literature on B2CVCs, specifically addressing gaps in the academic research by integrating measures of beliefs, attitudes, and behaviors in one model. The results provide additional insights to behavioral factors in a B2CVC environment, helping to sort out relationships between traditional measures and relatively new measures. For practitioners, the identification of factors, such as member involvement, that strongly influence B2CVC member satisfaction can help focus technological resources in key areas. Global e-marketers can develop marketing strategies directly targeting B2CVC members. In the global tourism business, they can target Chinese members of a B2CVC by providing special discounts for active community members or developing early adopter programs to encourage stickiness in the community. Future studies are called for, and more sophisticated modeling, to expand the measurement of B2CVC member behavior and to conduct experiments across industries, communities, and cultures.

• Journal of Intelligence and Information Systems
• /
• v.26 no.2
• /
• pp.79-104
• /
• 2020

Recently, as deep learning has attracted attention, the use of deep learning is being considered as a method for solving problems in various fields. In particular, deep learning is known to have excellent performance when applied to applying unstructured data such as text, sound and images, and many studies have proven its effectiveness. Owing to the remarkable development of text and image deep learning technology, interests in image captioning technology and its application is rapidly increasing. Image captioning is a technique that automatically generates relevant captions for a given image by handling both image comprehension and text generation simultaneously. In spite of the high entry barrier of image captioning that analysts should be able to process both image and text data, image captioning has established itself as one of the key fields in the A.I. research owing to its various applicability. In addition, many researches have been conducted to improve the performance of image captioning in various aspects. Recent researches attempt to create advanced captions that can not only describe an image accurately, but also convey the information contained in the image more sophisticatedly. Despite many recent efforts to improve the performance of image captioning, it is difficult to find any researches to interpret images from the perspective of domain experts in each field not from the perspective of the general public. Even for the same image, the part of interests may differ according to the professional field of the person who has encountered the image. Moreover, the way of interpreting and expressing the image also differs according to the level of expertise. The public tends to recognize the image from a holistic and general perspective, that is, from the perspective of identifying the image's constituent objects and their relationships. On the contrary, the domain experts tend to recognize the image by focusing on some specific elements necessary to interpret the given image based on their expertise. It implies that meaningful parts of an image are mutually different depending on viewers' perspective even for the same image. So, image captioning needs to implement this phenomenon. Therefore, in this study, we propose a method to generate captions specialized in each domain for the image by utilizing the expertise of experts in the corresponding domain. Specifically, after performing pre-training on a large amount of general data, the expertise in the field is transplanted through transfer-learning with a small amount of expertise data. However, simple adaption of transfer learning using expertise data may invoke another type of problems. Simultaneous learning with captions of various characteristics may invoke so-called 'inter-observation interference' problem, which make it difficult to perform pure learning of each characteristic point of view. For learning with vast amount of data, most of this interference is self-purified and has little impact on learning results. On the contrary, in the case of fine-tuning where learning is performed on a small amount of data, the impact of such interference on learning can be relatively large. To solve this problem, therefore, we propose a novel 'Character-Independent Transfer-learning' that performs transfer learning independently for each character. In order to confirm the feasibility of the proposed methodology, we performed experiments utilizing the results of pre-training on MSCOCO dataset which is comprised of 120,000 images and about 600,000 general captions. Additionally, according to the advice of an art therapist, about 300 pairs of 'image / expertise captions' were created, and the data was used for the experiments of expertise transplantation. As a result of the experiment, it was confirmed that the caption generated according to the proposed methodology generates captions from the perspective of implanted expertise whereas the caption generated through learning on general data contains a number of contents irrelevant to expertise interpretation. In this paper, we propose a novel approach of specialized image interpretation. To achieve this goal, we present a method to use transfer learning and generate captions specialized in the specific domain. In the future, by applying the proposed methodology to expertise transplant in various fields, we expected that many researches will be actively conducted to solve the problem of lack of expertise data and to improve performance of image captioning.

• Journal of the Korean Society of Groundwater Environment
• /
• v.7 no.3
• /
• pp.103-115
• /
• 2000

The formation of brown-colored precipitates is one of the serious problems frequently encountered in the development and supply of groundwater in Korea, because by it the water exceeds the drinking water standard in terms of color. taste. turbidity and dissolved iron concentration and of often results in scaling problem within the water supplying system. In groundwaters from the Pajoo area, brown precipitates are typically formed in a few hours after pumping-out. In this paper we examine the process of the brown precipitates' formation using the equilibrium thermodynamic and kinetic approaches, in order to understand the origin and geochemical pathway of the generation of turbidity in groundwater. The results of this study are used to suggest not only the proper pumping technique to minimize the formation of precipitates but also the optimal design of water treatment methods to improve the water quality. The bed-rock groundwater in the Pajoo area belongs to the Ca-$HCO_3$type that was evolved through water/rock (gneiss) interaction. Based on SEM-EDS and XRD analyses, the precipitates are identified as an amorphous, Fe-bearing oxides or hydroxides. By the use of multi-step filtration with pore sizes of 6, 4, 1, 0.45 and 0.2 $\mu\textrm{m}$, the precipitates mostly fall in the colloidal size (1 to 0.45 $\mu\textrm{m}$) but are concentrated (about 81%) in the range of 1 to 6 $\mu\textrm{m}$in teams of mass (weight) distribution. Large amounts of dissolved iron were possibly originated from dissolution of clinochlore in cataclasite which contains high amounts of Fe (up to 3 wt.%). The calculation of saturation index (using a computer code PHREEQC), as well as the examination of pH-Eh stability relations, also indicate that the final precipitates are Fe-oxy-hydroxide that is formed by the change of water chemistry (mainly, oxidation) due to the exposure to oxygen during the pumping-out of Fe(II)-bearing, reduced groundwater. After pumping-out, the groundwater shows the progressive decreases of pH, DO and alkalinity with elapsed time. However, turbidity increases and then decreases with time. The decrease of dissolved Fe concentration as a function of elapsed time after pumping-out is expressed as a regression equation Fe(II)=10.l exp(-0.0009t). The oxidation reaction due to the influx of free oxygen during the pumping and storage of groundwater results in the formation of brown precipitates, which is dependent on time, $Po_2$and pH. In order to obtain drinkable water quality, therefore, the precipitates should be removed by filtering after the stepwise storage and aeration in tanks with sufficient volume for sufficient time. Particle size distribution data also suggest that step-wise filtration would be cost-effective. To minimize the scaling within wells, the continued (if possible) pumping within the optimum pumping rate is recommended because this technique will be most effective for minimizing the mixing between deep Fe(II)-rich water and shallow $O_2$-rich water. The simultaneous pumping of shallow $O_2$-rich water in different wells is also recommended.

• Korean Journal of Breeding Science
• /
• v.51 no.3
• /
• pp.214-221
• /
• 2019

'Aram' is a soybean cultivar developed for soy-sprout. It was developed from the crossing of 'Bosug' (Glycine max IT213209) and 'Camp' (G. max IT267356) cultivars in 2007. F₁ plants and F₂ population were developed in 2009 and 2010. A promising line was selected in the F₅ generation in 2011 using the pedigree method and it was evaluated for agronomic traits, yield, and soy-sprouts characteristics in a preliminary yield trial (PYT) in 2012 and an advanced yield trial (AYT) in 2013. Agronomic traits and yield were stable between 2014 and 2016 in the regional yield trial (RYT) in four regions (Suwon, Naju, Dalseong, and Jeju). Morphological characteristics of 'Aram' are as follows: determinate plant type, purple flowers, grey pubescence, yellow pods, and small, yellow, and spherical seeds (9.9 g 100-seeds^-1) with a light brown hilum. The flowering date was the 5^th of August and the maturity date was the 15^th of October. Plant height, first pod height, number of nods, number of branches, and number of pods were 65 cm, 13 cm, 16, 4.5, and 99, respectively. In the sprout test, germination rate and sprout characteristics of 'Aram' were comparable to that of the 'Pungsannamulkong' cultivar. The yield of 'Aram' was 3.59 ton ha^-1 and it was 12% higher than that of 'Pungsannamulkong' in southern area of Korea. The yield of 'Aram' in the Jeju region, which is the main region for soybean sprout production, was 20% higher than that of 'Pungsannamulkong'. The height of the first pod and the tolerance to lodging and pod shattering, which are connected to the adaptation to mechanized harvesting, were higher in 'Aram' compared to those in 'Pungsannamulkong'. Therefore, the 'Aram' cultivar is expected to be broadly cultivated because of its higher soybean sprout quality, and seed yield and better adaptation to mechanized harvesting. (Registration number: 7718)

• Korean Journal of Breeding Science
• /
• v.51 no.4
• /
• pp.496-503
• /
• 2019

The soybean cultivar, 'Saegeum', has been developed for preparing soy-paste and tofu. The soybean cultivars 'Daepung' and 'SS98207-3SSD-168' were crossed in 2003 to obtain 'Saegeum'. Single seed descent method was used to advance the generation from F₃ to F₅, and the plant lines with promising traits were selected from F₆ to F₇ by pedigree method. The preliminary yield (PYT) and advanced yield trials (AYT) were conducted from 2009 to 2010, and the regional yield trial (RYT) was conducted in 12 regions between 2011 and 2013. The morphological characteristics of 'Saegeum' were as follows: determinate plant type, white flower, tawny pubescence color, and brown pod color. Flowering and maturity dates were August 2, XXXX and October 17, XXXX, respectively. Plant height, first pod height, number of nodes, number of branches, and number of pods were 79 cm, 18 cm, 16, 2.3, and 44, respectively. The seed characteristics of 'Saegeum' were as follows: yellow spherical shape, yellow hilum, and the 100-seed weight was 25.4 g. 'Saegeum' was resistant to bacterial pustule and SMV in the field test, and its lodging resistance was mildly strong, whereas its shattering resistance was excellent. The ability of this cultivar to be processed into tofu, soybean malt, and other fermented products was comparable with that of 'Daewonkong'. The yield of 'Saegeum' in the adaptable regions was 3.02 ton ha^-1. Thus, 'Saegeum' is adaptable to mechanized harvesting because of its high first pod height, as well as lodging and shattering resistance. (Registration number: 5929)

• Journal of Intelligence and Information Systems
• /
• v.26 no.4
• /
• pp.127-148
• /
• 2020

The data center is a physical environment facility for accommodating computer systems and related components, and is an essential foundation technology for next-generation core industries such as big data, smart factories, wearables, and smart homes. In particular, with the growth of cloud computing, the proportional expansion of the data center infrastructure is inevitable. Monitoring the health of these data center facilities is a way to maintain and manage the system and prevent failure. If a failure occurs in some elements of the facility, it may affect not only the relevant equipment but also other connected equipment, and may cause enormous damage. In particular, IT facilities are irregular due to interdependence and it is difficult to know the cause. In the previous study predicting failure in data center, failure was predicted by looking at a single server as a single state without assuming that the devices were mixed. Therefore, in this study, data center failures were classified into failures occurring inside the server (Outage A) and failures occurring outside the server (Outage B), and focused on analyzing complex failures occurring within the server. Server external failures include power, cooling, user errors, etc. Since such failures can be prevented in the early stages of data center facility construction, various solutions are being developed. On the other hand, the cause of the failure occurring in the server is difficult to determine, and adequate prevention has not yet been achieved. In particular, this is the reason why server failures do not occur singularly, cause other server failures, or receive something that causes failures from other servers. In other words, while the existing studies assumed that it was a single server that did not affect the servers and analyzed the failure, in this study, the failure occurred on the assumption that it had an effect between servers. In order to define the complex failure situation in the data center, failure history data for each equipment existing in the data center was used. There are four major failures considered in this study: Network Node Down, Server Down, Windows Activation Services Down, and Database Management System Service Down. The failures that occur for each device are sorted in chronological order, and when a failure occurs in a specific equipment, if a failure occurs in a specific equipment within 5 minutes from the time of occurrence, it is defined that the failure occurs simultaneously. After configuring the sequence for the devices that have failed at the same time, 5 devices that frequently occur simultaneously within the configured sequence were selected, and the case where the selected devices failed at the same time was confirmed through visualization. Since the server resource information collected for failure analysis is in units of time series and has flow, we used Long Short-term Memory (LSTM), a deep learning algorithm that can predict the next state through the previous state. In addition, unlike a single server, the Hierarchical Attention Network deep learning model structure was used in consideration of the fact that the level of multiple failures for each server is different. This algorithm is a method of increasing the prediction accuracy by giving weight to the server as the impact on the failure increases. The study began with defining the type of failure and selecting the analysis target. In the first experiment, the same collected data was assumed as a single server state and a multiple server state, and compared and analyzed. The second experiment improved the prediction accuracy in the case of a complex server by optimizing each server threshold. In the first experiment, which assumed each of a single server and multiple servers, in the case of a single server, it was predicted that three of the five servers did not have a failure even though the actual failure occurred. However, assuming multiple servers, all five servers were predicted to have failed. As a result of the experiment, the hypothesis that there is an effect between servers is proven. As a result of this study, it was confirmed that the prediction performance was superior when the multiple servers were assumed than when the single server was assumed. In particular, applying the Hierarchical Attention Network algorithm, assuming that the effects of each server will be different, played a role in improving the analysis effect. In addition, by applying a different threshold for each server, the prediction accuracy could be improved. This study showed that failures that are difficult to determine the cause can be predicted through historical data, and a model that can predict failures occurring in servers in data centers is presented. It is expected that the occurrence of disability can be prevented in advance using the results of this study.

• The Korean Journal of Mycology
• /
• v.7 no.1
• /
• pp.13-73
• /
• 1979

These studies were conducted to investigate nutrient sources and supplementary materials of synthetic compost media for Agaricus bisporus culture. Investigation were carried out to establish the optimum composition for compost of Agaricus bisporus methods of out-door fermentation and peakheating with rice straw as the main substrate of the media. The incidence and flora of harmful organisms in rice straw compost and their control were also studied. 1. When rice straw was used as the main substrate in synthetic compost as a carbon source. yields were remarkably high. Fermentation was more rapid than that of barley straw or wheat straw, and the total nitrogen content was high in rice straw compost. 2. Since the morphological and physico-chemical nature of Japonica and Indica types of rice straw are greatly dissimilar. there were apparent differences in the process of compost fermentation. Fermentation of Indica type straw proceeded more rapidly with a shortening the compost period, reducing the water supply, and required adding of supplementary materials for producing stable physical conditions. 3. Use of barley straw compost resulted in a smaller crop compared with rice straw. but when a 50%, barley straw and 50% rice straw mixture was used, the yield was almost the same as that using only rice straw. 4. There were extremely high positive correlations between yield of Agaricus bisporus and the total nitrogen, organic nitrogen, amino acids, amides and amino sugar nitrogen content of compost. The mycerial growth and fruit body formation were severely inhibited by ammonium nitrogen. 5. When rice straw was used as the main substrate for compost media, urea was the most suitable source of nitrogen. Poor results were obtained with calcium cyanamide and ammonium sulfate. When urea was applied three separate times, nitrogen loss during composting was decreased and the total nitrogen content of compost was increased. 6. The supplementation of organic nutrient activated compost fermentation and increased yield of Agaricus bisporus. The best sources of organic nutrients were: perilla meal, sesame meal, wheat bran and poultry manure, etc. 7. Soybean meal, tobacco powder and glutamic acid fermentation by-products which were industrial wastes, could be substituted for perilla meal, sesame meal and wheat bran as organic nutrient sources for compost media. B. When gypsum and zeolite were added to rice straw. physical deterioration of compost due to excess moisture and caramelization was observed. The Indica type of straw was more remarkable in increase of yield of Agricus bisporus by addition of supplementing materials than Japonica straw. 9. For preparing rice straw compost, the best mixture was prepared by 10% poultry manure, 5% perilla meal, 1. 2 to 1. 5% urea and 1% gypsum. At spring cropping, it was good to add rice bran to accelerate heat generation of the compost heap. 10. There was significantly high positive correlation (r=0.97) between accumulated temperature and the decomposition degree of compost during outdoor composting. The yield was highest at accumulated temperatures between 900 and $1,000^{\circ}C$. 11. Prolonging the composting period brought about an increase in decomposition degree and total nitrogen content, but a decrease in ammonium nitrogen. In the spring the suitable period of composting was 20 to 25 days. and about 15 days in autumn. For those periods, the degree of decomposition was 19 to 24%. 12. Compactness of wet compost at filling caused an increase in the residual ammonium nitrogen. methane and organic acid during peak heating. There was negative correlation between methane content and yield (r=0.76)and the same was true between volatile organic acid and yield (r=0.73). 13. In compost with a moisture content range between 69 to 80% at filling. the higher the moisture content, the lower the yield (r=0.78). This result was attributed to a reduction in the porosity of compost at filling the beds. The optimum porosity for good fermentation was between 41 and 53%. 14. Peak heating of the compost was essential for the prevention of harmful microorganisms and insect pests. and for the removal of excess ammonia. It was necessary to continue fer mentatiion for four days after peak heating. 15. Ten species of fungi which are harmful or competitive to Agaricus bisporus were identified from the rice compost, including Diehliomyces microsporus, Trichoderma sp. and Stysanus stemoites. The frequency of occurrance was notably high with serious damage to Agaricus bisporus. 16. Diehliomyces microsporus could be controlled by temperature adjustment of the growing room and by fumigating the compost and the house with Basamid and Vapam. Trichoderma was prevented by the use of Bavistin and Benomyl. 17. Four species of nematodes and five species of mites occured in compost during out-door composting. These orgnanisms could be controlled through peakheating compost for 6 hours at $60^{\circ}C$.

• Journal of Korean Society of Forest Science
• /
• v.54 no.1
• /
• pp.1-24
• /
• 1981

Kangweon-Do is rich in sightseeing resources. There are three sightseeing areas;first, mountain area including Seolak and Ohdae National Parks, and chiak Provincial Park; second eastern coastal area; third lake area including the watersheds of North Han River. In this paper, several methods of forestation were studied for landscaping the North Han River watersheds centering around Chounchon. In Chunchon lake complex, there are four lakes; Uiam, Chunchon, Soyang and Paro from down to upper stream. The total surface area of the above four lakes is $14.4km^2$ the total pondage of them 4,155 million $m^3$, the total generation of electric power of them 410 thousand Kw, and the total forest area bordering on them $1,208km^2$. The bordering forest consists of planned management forest ($745km^2$) and non-planned management forest ($463km^2$). The latter is divided into green belt zone, natural conservation area, and protection forest. The forest in green belt amounts to $177km^2$ and centers around the 10km radios from Chunchon. The forest in natural conservation area amounts to $165km^2$, which is established within 2km sight range from the Soyang-lake sides. Protection forest surrounding the lakes is $121km^2$ There are many scenic places, recreation gardens, cultural goods and ruins in this lake complex, which are the same good tourist resources as lakes and forest. The forest encirelng the lakes has the poor average growing stock of $15m^3/ha$, because 70% of the forest consists of the young plantation of 1 to 2 age class. The ration of the needle-leaved forest, the broad-leaved forest and the mixed forest in 35:37:28. From the standpoint of ownership, the forest consists of national forest (36%), provincial forest (14%), Gun forest (5%) and private forest(45%). The greater part of the forest soil, originated from granite and gneiss, is much liable to weathering. Because the surface soil is mostly sterile, the fertilization for improving the soil quality is strongly urged. Considering the above-mentioned, the forestation methods for improving landscape of the North Han River Watersheds are suggested as follows: 1) The mature-stage forest should be induced by means of fertilizing and tendering, as the forest in this area is the young plantation with poor soil. 2) The bare land should be afforested by planting the rapid growing species, such as rigida pine, alder, and etc. 3) The bare land in the canyon with moderate moist and comparatively rich soil should be planted with Korean-pine, larch, ro fir. 4) Japaness-pine stand should be changed into Korean-pine, fir, spruce or hemlock stand from ravine to top gradually, because the Japanese-pine has poor capacity of water conservation and great liability to pine gall midge. 5) Present hard-wood forest, consisting of miscellaneous trees comparatively less valuable from the point of wood quality and scenerity, should be change into oak, maple, fraxinus-rhynchophylla, birch or juglan stand which is comparatively more valuable. 6) In the mountain foot within the sight-range, stands should be established with such species as cherry, weeping willow, white poplar, machilus, maiden-hair tree, juniper, chestnut or apricot. 7) The regeneration of some broad-leaved forests should be induced to the middle forest type, leading to the harmonious arrangement of the two storied forest and the coppice. 8) For the preservation of scenery, the reproduction of the soft-wood forest should be done under the selection method or the shelter-wood system. 9) Mixed forest should be regenerated under the middle forest system with upper needle-leaved forest and lower broad-leaved forest. In brief, the nature's mysteriousness should be conserved by combining the womanly elegance of the lakes and the manly grandeur of the forest.