• Journal of Intelligence and Information Systems
• /
• v.24 no.3
• /
• pp.1-19
• /
• 2018

Large amount of data is now available for research and business sectors to extract knowledge from it. This data can be in the form of unstructured data such as audio, text, and image data and can be analyzed by deep learning methodology. Deep learning is now widely used for various estimation, classification, and prediction problems. Especially, fashion business adopts deep learning techniques for apparel recognition, apparel search and retrieval engine, and automatic product recommendation. The core model of these applications is the image classification using Convolutional Neural Networks (CNN). CNN is made up of neurons which learn parameters such as weights while inputs come through and reach outputs. CNN has layer structure which is best suited for image classification as it is comprised of convolutional layer for generating feature maps, pooling layer for reducing the dimensionality of feature maps, and fully-connected layer for classifying the extracted features. However, most of the classification models have been trained using online product image, which is taken under controlled situation such as apparel image itself or professional model wearing apparel. This image may not be an effective way to train the classification model considering the situation when one might want to classify street fashion image or walking image, which is taken in uncontrolled situation and involves people's movement and unexpected pose. Therefore, we propose to train the model with runway apparel image dataset which captures mobility. This will allow the classification model to be trained with far more variable data and enhance the adaptation with diverse query image. To achieve both convergence and generalization of the model, we apply Transfer Learning on our training network. As Transfer Learning in CNN is composed of pre-training and fine-tuning stages, we divide the training step into two. First, we pre-train our architecture with large-scale dataset, ImageNet dataset, which consists of 1.2 million images with 1000 categories including animals, plants, activities, materials, instrumentations, scenes, and foods. We use GoogLeNet for our main architecture as it has achieved great accuracy with efficiency in ImageNet Large Scale Visual Recognition Challenge (ILSVRC). Second, we fine-tune the network with our own runway image dataset. For the runway image dataset, we could not find any previously and publicly made dataset, so we collect the dataset from Google Image Search attaining 2426 images of 32 major fashion brands including Anna Molinari, Balenciaga, Balmain, Brioni, Burberry, Celine, Chanel, Chloe, Christian Dior, Cividini, Dolce and Gabbana, Emilio Pucci, Ermenegildo, Fendi, Giuliana Teso, Gucci, Issey Miyake, Kenzo, Leonard, Louis Vuitton, Marc Jacobs, Marni, Max Mara, Missoni, Moschino, Ralph Lauren, Roberto Cavalli, Sonia Rykiel, Stella McCartney, Valentino, Versace, and Yve Saint Laurent. We perform 10-folded experiments to consider the random generation of training data, and our proposed model has achieved accuracy of 67.2% on final test. Our research suggests several advantages over previous related studies as to our best knowledge, there haven't been any previous studies which trained the network for apparel image classification based on runway image dataset. We suggest the idea of training model with image capturing all the possible postures, which is denoted as mobility, by using our own runway apparel image dataset. Moreover, by applying Transfer Learning and using checkpoint and parameters provided by Tensorflow Slim, we could save time spent on training the classification model as taking 6 minutes per experiment to train the classifier. This model can be used in many business applications where the query image can be runway image, product image, or street fashion image. To be specific, runway query image can be used for mobile application service during fashion week to facilitate brand search, street style query image can be classified during fashion editorial task to classify and label the brand or style, and website query image can be processed by e-commerce multi-complex service providing item information or recommending similar item.

• Tuberculosis and Respiratory Diseases
• /
• v.55 no.2
• /
• pp.140-153
• /
• 2003

Background : The cell-mediated immune reaction to tuberculosis infection involves a complex network of cytokines. The extent of inflammation, tissue damage and severity of the disease suggested to be determined by the balance between extent and duration of the proinflammatory cytokine response versus those of the suppressive cytokines. The systemic cytokine response in pathogenesis of tuberculosis can be assessed by measuring serum cytokine levels. Method : Serum interleukin-1 beta(IL-$1{\beta}$), IL-2, IL-4, IL-6, IL-10, IL-12(p40), tumor necrosis factor-alpha(TNF-${\alpha}$), interferon-gamma(IFN-${\gamma}$) and transforming growth factor-beta(TGF-${\beta}$) levels were measured in 83 patients with pulmonary tuberculosis, 10 patients with endobronchial tuberculosis before treatment and 20 healthy subjects by using a sandwich ELISA. In patients with pulmonary tuberculosis, they were divided into mild, moderate and far advanced group according to the severity by ATS guidelines. To compare with those of pretreatment levels, we measured serum IL-$1{\beta}$, IL-2, IL-4, IL-6, IL-10, IL-12(p40), TNF-${\alpha}$, IFN-${\gamma}$ and TGF-${\beta}$ levels in 45 of 83 patients with pulmonary tuberculosis after 2 and 6 months of treatment. Results : 1) In sera of patients with active pulmonary tuberculosis(n=83), IL-$1{\beta}$, IL-6(p<0.05), TNF-${\alpha}$, and IFN-${\gamma}$ were elevated and TGF-${\beta}$ was decreased comparing to control. IL-2, Il-12(p40), IL-4 and IL-10 were similar between the patients with tuberculosis and control. 2) In endobronchial tuberculosis, IL-6 and TNF-${\alpha}$ were elevated and TGF-${\beta}$ was decreased comparing to control. IL-12(p40) seemed to be elevated comparing to pulmonary tuberculosis. 3) Far advanced tuberculosis showed markedly elevated IL-6 and IFN-${\gamma}$ level(p<0.05). 4) The significant correlations were noted between IL-1, IL-6 AND TNF-${\alpha}$ and between IL-12, Il-2 and IL-4(p<0.01). 5) After 2 and 6 months of standard treatment, the level of IL-6 and IFN-${\gamma}$ was significantly decreased(p<0.05). Conclusion : These results showed that an altered balance between cytokines is likely to be involved in the extent of inflammation, tissue damage and severity of the disease tuberculosis. But, it should be considered diversities of cytokine response according to type of tuberculosis and immunity in clinical application and interpreting future studies.

• Journal of Distribution Research
• /
• v.17 no.1
• /
• pp.37-63
• /
• 2012

In franchise business, exclusive sales territory (sometimes EST in table) protection is a very important issue from an economic, social and political point of view. It affects the growth and survival of both franchisor and franchisee and often raises issues of social and political conflicts. When franchisee is not familiar with related laws and regulations, franchisor has high chance to utilize it. Exclusive sales territory protection by the manufacturer and distributors (wholesalers or retailers) means sales area restriction by which only certain distributors have right to sell products or services. The distributor, who has been granted exclusive sales territories, can protect its own territory, whereas he may be prohibited from entering in other regions. Even though exclusive sales territory is a quite critical problem in franchise business, there is not much rigorous research about the reason, results, evaluation, and future direction based on empirical data. This paper tries to address this problem not only from logical and nomological validity, but from empirical validation. While we purse an empirical analysis, we take into account the difficulties of real data collection and statistical analysis techniques. We use a set of disclosure document data collected by Korea Fair Trade Commission, instead of conventional survey method which is usually criticized for its measurement error. Existing theories about exclusive sales territory can be summarized into two groups as shown in the table below. The first one is about the effectiveness of exclusive sales territory from both franchisor and franchisee point of view. In fact, output of exclusive sales territory can be positive for franchisors but negative for franchisees. Also, it can be positive in terms of sales but negative in terms of profit. Therefore, variables and viewpoints should be set properly. The other one is about the motive or reason why exclusive sales territory is protected. The reasons can be classified into four groups - industry characteristics, franchise systems characteristics, capability to maintain exclusive sales territory, and strategic decision. Within four groups of reasons, there are more specific variables and theories as below. Based on these theories, we develop nine hypotheses which are briefly shown in the last table below with the results. In order to validate the hypothesis, data is collected from government (FTC) homepage which is open source. The sample consists of 1,896 franchisors and it contains about three year operation data, from 2006 to 2008. Within the samples, 627 have exclusive sales territory protection policy and the one with exclusive sales territory policy is not evenly distributed over 19 representative industries. Additional data are also collected from another government agency homepage, like Statistics Korea. Also, we combine data from various secondary sources to create meaningful variables as shown in the table below. All variables are dichotomized by mean or median split if they are not inherently dichotomized by its definition, since each hypothesis is composed by multiple variables and there is no solid statistical technique to incorporate all these conditions to test the hypotheses. This paper uses a simple chi-square test because hypotheses and theories are built upon quite specific conditions such as industry type, economic condition, company history and various strategic purposes. It is almost impossible to find all those samples to satisfy them and it can't be manipulated in experimental settings. However, more advanced statistical techniques are very good on clean data without exogenous variables, but not good with real complex data. The chi-square test is applied in a way that samples are grouped into four with two criteria, whether they use exclusive sales territory protection or not, and whether they satisfy conditions of each hypothesis. So the proportion of sample franchisors which satisfy conditions and protect exclusive sales territory, does significantly exceed the proportion of samples that satisfy condition and do not protect. In fact, chi-square test is equivalent with the Poisson regression which allows more flexible application. As results, only three hypotheses are accepted. When attitude toward the risk is high so loyalty fee is determined according to sales performance, EST protection makes poor results as expected. And when franchisor protects EST in order to recruit franchisee easily, EST protection makes better results. Also, when EST protection is to improve the efficiency of franchise system as a whole, it shows better performances. High efficiency is achieved as EST prohibits the free riding of franchisee who exploits other's marketing efforts, and it encourages proper investments and distributes franchisee into multiple regions evenly. Other hypotheses are not supported in the results of significance testing. Exclusive sales territory should be protected from proper motives and administered for mutual benefits. Legal restrictions driven by the government agency like FTC could be misused and cause mis-understandings. So there need more careful monitoring on real practices and more rigorous studies by both academicians and practitioners.

• Journal of the Korean Institute of Traditional Landscape Architecture
• /
• v.31 no.2
• /
• pp.43-57
• /
• 2013

The purpose of this study is to identify potential semantic landscape makeup of "the Forty Eight Poems of Soswaewon" according to Yin-Yang and Five Elements Principle(陰陽五行論). that speculation system between human's nature and cosmical universal order. Existing academic discussions made so far concerning this topic can be summed up as follows: 1. Among Yin-Yang-based landscape makeups of the Forty Eight Poems of Soswaewon, poetic writings for embodiment of interactions between nature and human behaviors focused on depicting dynamic aspects of a poetic narrator when he appreciates or explores hills and streams as of to live free from worldly cares. Primarily, many of those writings were created on the east and south primarily through assignment of yang. On the other hand, poetic writings for embodiment of nature and seasonal scenery - as static landscape makeup of yin - were often created on or near the north and west for many times. Those writings focusing on embodiment of nature and artificial scenery as a work are divided into two categories: One category refers to author Kim In-hu's expression of semantic landscape from seasonal scenery in nature. The other refers to his depiction of realistic garden images as they are. In the Forty Eight Poems of Soswaewon, the poetic writings show that author Kim focused on embodying seasonal scenery rather than expressing human behaviors. In addition, both Poem No. 1 and Poem No. 48(last poem; titled 'Jangwon Jeyeong') were created in a same place, which author Kim sought to understand the place as a space of beginning and end where yin and yang - i.e. the principle of natural cycle - are inherent. 2. According to construction about landscape in the Forty Eight Poems of Soswaewon on the basis of Ohaeng-ron (five natural element principle), it was found that tree(木) and fire(火) are typical examples of a world combined by emanation. First, many of poetic writings depicting the sentiments of tree focused on embodying seasonal scenery and were located in the place of Ogogmun(五曲門) area in the east, from overall perspective of Soswaewon. The content of these poems shows generation and curve / straightness in flexibility and simplicity. Many of poems depicting the sentiments of fire(火) focused on embodying human behaviors, and they were created in Aeyangdan area on the south of Soswaewon over which sun rises at noon. These poems are all on a status of side movement that is characterized by emanation and ascension which belong to attributes of yang. 3. With regard to Ohaeng-ron's interpretation about landscape in the Forty Eight Poems of Soswaewon, it was found that metal(金) and water(水) are typical examples of world combined by convergence. First, it was found that all of poems depicting sentiments of metal focused on embodying seasonal scenery, and were created in a bamboo grove area on the west from overall perspective of Soswaewon. They represent scenery of autumn among 4 seasons to symbolize faithfulness vested in a man of virtue(seonbi) with integrity and righteousness. Poems depicting sentiments of water were created in vicinity of Jewoldang on the north, possibly topmost of Soswaewon. They were divided into two categories: One category refers to poems embodying actions of welcoming the first full moon deep in the night after sunset, and the other refers to poems embodying natural scenery of snowscape. All of those poems focused on expressing any atmosphere of turning into yin via convergence. 4. With regard to Ohaeng-ron's interpretation of landscape in the Forty Eight Poems of Soswaewon, it was found that poems depicting sentiments of earth(土), a complex body of convergence and emanation, were created in vicinity of mountain stream around Gwangpunggak which is located in the center of Soswaewon. These poems focused on carrying actions of author Kim by way of natural phenomena and artificial scenery.

• Proceedings of the Korean DIstribution Association Conference
• /
• 2005.05a
• /
• pp.383-405
• /
• 2005

The traditional market is foremost a regionally positioned place, wherein the market directly represents regional and cultural centered traits while it plays an important role in the circulation of facilities through reciprocal, informative and cultural exchanges while sewing to form local communities. The traditional market in Korea is one of representative retail businesses and premodern marketing techniques by family owned business of less than five members such as product management, purchase method, and marketing patterns etc. Since the 1990s, the appearance of new circulation-type businesses and large discount convenience stores escalated the loss of traditional competitiveness, increased the living standard of customers, changed purchasing patterns, and expanded the ubiquity of the Internet. All of these changes in external circulation circumstances have led the traditional markets to lose their place in the economy. The traditional market should revive on a regional site basis through the formation of a community of regional neighbors and through knowledge-sharing that leads to the creation of wealth. For the purpose of creating a wealth in a place, the following components are necessary: 1) a facility suitable for the spatial place of the present, 2)trust built through exchanges within the changing market environment, which would simultaneously satisfy customer's desires, 3) international bench marking on cases such as regionally centered TCM (England), BID (USA), and TMO (Japan) so that the market unit of store placement transfers from a spot policy to a line policy, 4)conversion of communicative conception through a surface policy approach centered around a macro-region perspective. The budget of the traditional market funding policy was operational between 2001 and 2004, serving as a counter move to solve the problem of the old traditional market through government intervention in regional economies to promote national economic strength. This national treasury funding project was centered on environmental improvement, research corps, and business modernization through the expenditure of 3,853 hundred million won (Korean currency). However, the effectiveness of this project has yet to be to proven through investigation. Furthermore, in promoting this funding support project, a lack of professionalism among merchants in the market led to constant limitations in comprehensive striving strategies, reduced capabilities in middle-and long-term plan setup, and created reductions in voluntary merchant agreement solutions. The traditional market should go beyond mere physical place and ordinary products creative site strategies employing the communicative approach must accompany these strategies to make the market a new regional and spatial living place. Thus, regarding recent paradigm changes and the introduction of region-based site functions into the traditional market, acquiring a conversion of direction into the newly developed project is essential to reinvestigate the traditional market composed of cultural and economic meanings, for the purpose of the research. Excavating social policy demands through the comparative analysis of domestic and international cases as well as innovative and expert management leadership development for NPO or NGO civil entrepreneurs through advanced case research on present promotion methods is extremely important. Discovering the seeds of the cultural contents industry cored around regional resource usages, commercializing regionally reknowned products, and constructing complex cultural living places for regional networks are especially important. In order to accelerate these solutions, a comprehensive and systemized approach research operated within a mentor academy system is required, as research will reveal distinctive traits of the traditional market in the aging society.

• Radiation Oncology Journal
• /
• v.19 no.1
• /
• pp.53-65
• /
• 2001

Purpose : To improve the local control of patients with nasopharyngeal cancer, we have implemented 3-D conformal radiotherapy and forward intensity modulated radiation therapy (IMRT) to used of compensating filters. Three dimension conformal radiotherapy with intensity modulation is a new modality for cancer treatments. We designed 3-D treatment planning with 3-D RTP (radiation treatment planning system) and evaluation dose distribution with tumor control probability (TCP) and normal tissue complication probability (NTCP). Material and Methods : We have developed a treatment plan consisting four intensity modulated photon fields that are delivered through the compensating tilters and block transmission for critical organs. We get a full size CT imaging including head and neck as 3 mm slices, and delineating PTV (planning target volume) and surrounding critical organs, and reconstructed 3D imaging on the computer windows. In the planning stage, the planner specifies the number of beams and their directions including non-coplanar, and the prescribed doses for the target volume and the permissible dose of normal organs and the overlap regions. We designed compensating filter according to tissue deficit and PTV volume shape also dose weighting for each field to obtain adequate dose distribution, and shielding blocks weighting for transmission. Therapeutic gains were evaluated by numerical equation of tumor control probability and normal tissue complication probability. The TCP and NTCP by DVH (dose volume histogram) were compared with the 3-D conformal radiotherapy and forward intensity modulated conformal radiotherapy by compensator and blocks weighting. Optimization for the weight distribution was peformed iteration with initial guess weight or the even weight distribution. The TCP and NTCP by DVH were compared with the 3-D conformal radiotherapy and intensitiy modulated conformal radiotherapy by compensator and blocks weighting. Results : Using a four field IMRT plan, we have customized dose distribution to conform and deliver sufficient dose to the PTV. In addition, in the overlap regions between the PTV and the normal organs (spinal cord, salivary grand, pituitary, optic nerves), the dose is kept within the tolerance of the respective organs. We evaluated to obtain sufficient TCP value and acceptable NTCP using compensating filters. Quality assurance checks show acceptable agreement between the planned and the implemented MLC(multi-leaf collimator). Conclusion : IMRT provides a powerful and efficient solution for complex planning problems where the surrounding normal tissues place severe constraints on the prescription dose. The intensity modulated fields can be efficaciously and accurately delivered using compensating filters.

• Journal of Intelligence and Information Systems
• /
• v.19 no.4
• /
• pp.123-132
• /
• 2013

As the smartphones are equipped with various sensors such as the accelerometer, GPS, gravity sensor, gyros, ambient light sensor, proximity sensor, and so on, there have been many research works on making use of these sensors to create valuable applications. Human activity recognition is one such application that is motivated by various welfare applications such as the support for the elderly, measurement of calorie consumption, analysis of lifestyles, analysis of exercise patterns, and so on. One of the challenges faced when using the smartphone sensors for activity recognition is that the number of sensors used should be minimized to save the battery power. When the number of sensors used are restricted, it is difficult to realize a highly accurate activity recognizer or a classifier because it is hard to distinguish between subtly different activities relying on only limited information. The difficulty gets especially severe when the number of different activity classes to be distinguished is very large. In this paper, we show that a fairly accurate classifier can be built that can distinguish ten different activities by using only a single sensor data, i.e., the smartphone accelerometer data. The approach that we take to dealing with this ten-class problem is to use the ensemble of nested dichotomy (END) method that transforms a multi-class problem into multiple two-class problems. END builds a committee of binary classifiers in a nested fashion using a binary tree. At the root of the binary tree, the set of all the classes are split into two subsets of classes by using a binary classifier. At a child node of the tree, a subset of classes is again split into two smaller subsets by using another binary classifier. Continuing in this way, we can obtain a binary tree where each leaf node contains a single class. This binary tree can be viewed as a nested dichotomy that can make multi-class predictions. Depending on how a set of classes are split into two subsets at each node, the final tree that we obtain can be different. Since there can be some classes that are correlated, a particular tree may perform better than the others. However, we can hardly identify the best tree without deep domain knowledge. The END method copes with this problem by building multiple dichotomy trees randomly during learning, and then combining the predictions made by each tree during classification. The END method is generally known to perform well even when the base learner is unable to model complex decision boundaries As the base classifier at each node of the dichotomy, we have used another ensemble classifier called the random forest. A random forest is built by repeatedly generating a decision tree each time with a different random subset of features using a bootstrap sample. By combining bagging with random feature subset selection, a random forest enjoys the advantage of having more diverse ensemble members than a simple bagging. As an overall result, our ensemble of nested dichotomy can actually be seen as a committee of committees of decision trees that can deal with a multi-class problem with high accuracy. The ten classes of activities that we distinguish in this paper are 'Sitting', 'Standing', 'Walking', 'Running', 'Walking Uphill', 'Walking Downhill', 'Running Uphill', 'Running Downhill', 'Falling', and 'Hobbling'. The features used for classifying these activities include not only the magnitude of acceleration vector at each time point but also the maximum, the minimum, and the standard deviation of vector magnitude within a time window of the last 2 seconds, etc. For experiments to compare the performance of END with those of other methods, the accelerometer data has been collected at every 0.1 second for 2 minutes for each activity from 5 volunteers. Among these 5,900 ($=5{\times}(60{\times}2-2)/0.1$) data collected for each activity (the data for the first 2 seconds are trashed because they do not have time window data), 4,700 have been used for training and the rest for testing. Although 'Walking Uphill' is often confused with some other similar activities, END has been found to classify all of the ten activities with a fairly high accuracy of 98.4%. On the other hand, the accuracies achieved by a decision tree, a k-nearest neighbor, and a one-versus-rest support vector machine have been observed as 97.6%, 96.5%, and 97.6%, respectively.

• Journal of Intelligence and Information Systems
• /
• v.20 no.3
• /
• pp.19-43
• /
• 2014

The Pareto principle, also known as the 80-20 rule, states that roughly 80% of the effects come from 20% of the causes for many events including natural phenomena. It has been recognized as a golden rule in business with a wide application of such discovery like 20 percent of customers resulting in 80 percent of total sales. On the other hand, the Long Tail theory, pointing out that "the trivial many" produces more value than "the vital few," has gained popularity in recent times with a tremendous reduction of distribution and inventory costs through the development of ICT(Information and Communication Technology). This study started with a view to illuminating how these two primary business paradigms-Pareto principle and Long Tail theory-relates to the success of virtual knowledge collaboration. The importance of virtual knowledge collaboration is soaring in this era of globalization and virtualization transcending geographical and temporal constraints. Many previous studies on knowledge sharing have focused on the factors to affect knowledge sharing, seeking to boost individual knowledge sharing and resolve the social dilemma caused from the fact that rational individuals are likely to rather consume than contribute knowledge. Knowledge collaboration can be defined as the creation of knowledge by not only sharing knowledge, but also by transforming and integrating such knowledge. In this perspective of knowledge collaboration, the relative distribution of knowledge sharing among participants can count as much as the absolute amounts of individual knowledge sharing. In particular, whether the more contribution of the upper 20 percent of participants in knowledge sharing will enhance the efficiency of overall knowledge collaboration is an issue of interest. This study deals with the effect of this sort of knowledge sharing distribution on the efficiency of knowledge collaboration and is extended to reflect the work characteristics. All analyses were conducted based on actual data instead of self-reported questionnaire surveys. More specifically, we analyzed the collaborative behaviors of editors of 2,978 English Wikipedia featured articles, which are the best quality grade of articles in English Wikipedia. We adopted Pareto ratio, the ratio of the number of knowledge contribution of the upper 20 percent of participants to the total number of knowledge contribution made by the total participants of an article group, to examine the effect of Pareto principle. In addition, Gini coefficient, which represents the inequality of income among a group of people, was applied to reveal the effect of inequality of knowledge contribution. Hypotheses were set up based on the assumption that the higher ratio of knowledge contribution by more highly motivated participants will lead to the higher collaboration efficiency, but if the ratio gets too high, the collaboration efficiency will be exacerbated because overall informational diversity is threatened and knowledge contribution of less motivated participants is intimidated. Cox regression models were formulated for each of the focal variables-Pareto ratio and Gini coefficient-with seven control variables such as the number of editors involved in an article, the average time length between successive edits of an article, the number of sections a featured article has, etc. The dependent variable of the Cox models is the time spent from article initiation to promotion to the featured article level, indicating the efficiency of knowledge collaboration. To examine whether the effects of the focal variables vary depending on the characteristics of a group task, we classified 2,978 featured articles into two categories: Academic and Non-academic. Academic articles refer to at least one paper published at an SCI, SSCI, A&HCI, or SCIE journal. We assumed that academic articles are more complex, entail more information processing and problem solving, and thus require more skill variety and expertise. The analysis results indicate the followings; First, Pareto ratio and inequality of knowledge sharing relates in a curvilinear fashion to the collaboration efficiency in an online community, promoting it to an optimal point and undermining it thereafter. Second, the curvilinear effect of Pareto ratio and inequality of knowledge sharing on the collaboration efficiency is more sensitive with a more academic task in an online community.

• Journal of Intelligence and Information Systems
• /
• v.25 no.4
• /
• pp.141-154
• /
• 2019

Rapid growth of internet technology and social media is progressing. Data mining technology has evolved to enable unstructured document representations in a variety of applications. Sentiment analysis is an important technology that can distinguish poor or high-quality content through text data of products, and it has proliferated during text mining. Sentiment analysis mainly analyzes people's opinions in text data by assigning predefined data categories as positive and negative. This has been studied in various directions in terms of accuracy from simple rule-based to dictionary-based approaches using predefined labels. In fact, sentiment analysis is one of the most active researches in natural language processing and is widely studied in text mining. When real online reviews aren't available for others, it's not only easy to openly collect information, but it also affects your business. In marketing, real-world information from customers is gathered on websites, not surveys. Depending on whether the website's posts are positive or negative, the customer response is reflected in the sales and tries to identify the information. However, many reviews on a website are not always good, and difficult to identify. The earlier studies in this research area used the reviews data of the Amazon.com shopping mal, but the research data used in the recent studies uses the data for stock market trends, blogs, news articles, weather forecasts, IMDB, and facebook etc. However, the lack of accuracy is recognized because sentiment calculations are changed according to the subject, paragraph, sentiment lexicon direction, and sentence strength. This study aims to classify the polarity analysis of sentiment analysis into positive and negative categories and increase the prediction accuracy of the polarity analysis using the pretrained IMDB review data set. First, the text classification algorithm related to sentiment analysis adopts the popular machine learning algorithms such as NB (naive bayes), SVM (support vector machines), XGboost, RF (random forests), and Gradient Boost as comparative models. Second, deep learning has demonstrated discriminative features that can extract complex features of data. Representative algorithms are CNN (convolution neural networks), RNN (recurrent neural networks), LSTM (long-short term memory). CNN can be used similarly to BoW when processing a sentence in vector format, but does not consider sequential data attributes. RNN can handle well in order because it takes into account the time information of the data, but there is a long-term dependency on memory. To solve the problem of long-term dependence, LSTM is used. For the comparison, CNN and LSTM were chosen as simple deep learning models. In addition to classical machine learning algorithms, CNN, LSTM, and the integrated models were analyzed. Although there are many parameters for the algorithms, we examined the relationship between numerical value and precision to find the optimal combination. And, we tried to figure out how the models work well for sentiment analysis and how these models work. This study proposes integrated CNN and LSTM algorithms to extract the positive and negative features of text analysis. The reasons for mixing these two algorithms are as follows. CNN can extract features for the classification automatically by applying convolution layer and massively parallel processing. LSTM is not capable of highly parallel processing. Like faucets, the LSTM has input, output, and forget gates that can be moved and controlled at a desired time. These gates have the advantage of placing memory blocks on hidden nodes. The memory block of the LSTM may not store all the data, but it can solve the CNN's long-term dependency problem. Furthermore, when LSTM is used in CNN's pooling layer, it has an end-to-end structure, so that spatial and temporal features can be designed simultaneously. In combination with CNN-LSTM, 90.33% accuracy was measured. This is slower than CNN, but faster than LSTM. The presented model was more accurate than other models. In addition, each word embedding layer can be improved when training the kernel step by step. CNN-LSTM can improve the weakness of each model, and there is an advantage of improving the learning by layer using the end-to-end structure of LSTM. Based on these reasons, this study tries to enhance the classification accuracy of movie reviews using the integrated CNN-LSTM model.

• Journal of Intelligence and Information Systems
• /
• v.26 no.2
• /
• pp.105-129
• /
• 2020

This study uses corporate data from 2012 to 2018 when K-IFRS was applied in earnest to predict default risks. The data used in the analysis totaled 10,545 rows, consisting of 160 columns including 38 in the statement of financial position, 26 in the statement of comprehensive income, 11 in the statement of cash flows, and 76 in the index of financial ratios. Unlike most previous prior studies used the default event as the basis for learning about default risk, this study calculated default risk using the market capitalization and stock price volatility of each company based on the Merton model. Through this, it was able to solve the problem of data imbalance due to the scarcity of default events, which had been pointed out as the limitation of the existing methodology, and the problem of reflecting the difference in default risk that exists within ordinary companies. Because learning was conducted only by using corporate information available to unlisted companies, default risks of unlisted companies without stock price information can be appropriately derived. Through this, it can provide stable default risk assessment services to unlisted companies that are difficult to determine proper default risk with traditional credit rating models such as small and medium-sized companies and startups. Although there has been an active study of predicting corporate default risks using machine learning recently, model bias issues exist because most studies are making predictions based on a single model. Stable and reliable valuation methodology is required for the calculation of default risk, given that the entity's default risk information is very widely utilized in the market and the sensitivity to the difference in default risk is high. Also, Strict standards are also required for methods of calculation. The credit rating method stipulated by the Financial Services Commission in the Financial Investment Regulations calls for the preparation of evaluation methods, including verification of the adequacy of evaluation methods, in consideration of past statistical data and experiences on credit ratings and changes in future market conditions. This study allowed the reduction of individual models' bias by utilizing stacking ensemble techniques that synthesize various machine learning models. This allows us to capture complex nonlinear relationships between default risk and various corporate information and maximize the advantages of machine learning-based default risk prediction models that take less time to calculate. To calculate forecasts by sub model to be used as input data for the Stacking Ensemble model, training data were divided into seven pieces, and sub-models were trained in a divided set to produce forecasts. To compare the predictive power of the Stacking Ensemble model, Random Forest, MLP, and CNN models were trained with full training data, then the predictive power of each model was verified on the test set. The analysis showed that the Stacking Ensemble model exceeded the predictive power of the Random Forest model, which had the best performance on a single model. Next, to check for statistically significant differences between the Stacking Ensemble model and the forecasts for each individual model, the Pair between the Stacking Ensemble model and each individual model was constructed. Because the results of the Shapiro-wilk normality test also showed that all Pair did not follow normality, Using the nonparametric method wilcoxon rank sum test, we checked whether the two model forecasts that make up the Pair showed statistically significant differences. The analysis showed that the forecasts of the Staging Ensemble model showed statistically significant differences from those of the MLP model and CNN model. In addition, this study can provide a methodology that allows existing credit rating agencies to apply machine learning-based bankruptcy risk prediction methodologies, given that traditional credit rating models can also be reflected as sub-models to calculate the final default probability. Also, the Stacking Ensemble techniques proposed in this study can help design to meet the requirements of the Financial Investment Business Regulations through the combination of various sub-models. We hope that this research will be used as a resource to increase practical use by overcoming and improving the limitations of existing machine learning-based models.