• Journal of Intelligence and Information Systems
• /
• v.22 no.1
• /
• pp.139-157
• /
• 2016

Bankruptcy involves considerable costs, so it can have significant effects on a country's economy. Thus, bankruptcy prediction is an important issue. Over the past several decades, many researchers have addressed topics associated with bankruptcy prediction. Early research on bankruptcy prediction employed conventional statistical methods such as univariate analysis, discriminant analysis, multiple regression, and logistic regression. Later on, many studies began utilizing artificial intelligence techniques such as inductive learning, neural networks, and case-based reasoning. Currently, ensemble models are being utilized to enhance the accuracy of bankruptcy prediction. Ensemble classification involves combining multiple classifiers to obtain more accurate predictions than those obtained using individual models. Ensemble learning techniques are known to be very useful for improving the generalization ability of the classifier. Base classifiers in the ensemble must be as accurate and diverse as possible in order to enhance the generalization ability of an ensemble model. Commonly used methods for constructing ensemble classifiers include bagging, boosting, and random subspace. The random subspace method selects a random feature subset for each classifier from the original feature space to diversify the base classifiers of an ensemble. Each ensemble member is trained by a randomly chosen feature subspace from the original feature set, and predictions from each ensemble member are combined by an aggregation method. The k-nearest neighbors (KNN) classifier is robust with respect to variations in the dataset but is very sensitive to changes in the feature space. For this reason, KNN is a good classifier for the random subspace method. The KNN random subspace ensemble model has been shown to be very effective for improving an individual KNN model. The k parameter of KNN base classifiers and selected feature subsets for base classifiers play an important role in determining the performance of the KNN ensemble model. However, few studies have focused on optimizing the k parameter and feature subsets of base classifiers in the ensemble. This study proposed a new ensemble method that improves upon the performance KNN ensemble model by optimizing both k parameters and feature subsets of base classifiers. A genetic algorithm was used to optimize the KNN ensemble model and improve the prediction accuracy of the ensemble model. The proposed model was applied to a bankruptcy prediction problem by using a real dataset from Korean companies. The research data included 1800 externally non-audited firms that filed for bankruptcy (900 cases) or non-bankruptcy (900 cases). Initially, the dataset consisted of 134 financial ratios. Prior to the experiments, 75 financial ratios were selected based on an independent sample t-test of each financial ratio as an input variable and bankruptcy or non-bankruptcy as an output variable. Of these, 24 financial ratios were selected by using a logistic regression backward feature selection method. The complete dataset was separated into two parts: training and validation. The training dataset was further divided into two portions: one for the training model and the other to avoid overfitting. The prediction accuracy against this dataset was used to determine the fitness value in order to avoid overfitting. The validation dataset was used to evaluate the effectiveness of the final model. A 10-fold cross-validation was implemented to compare the performances of the proposed model and other models. To evaluate the effectiveness of the proposed model, the classification accuracy of the proposed model was compared with that of other models. The Q-statistic values and average classification accuracies of base classifiers were investigated. The experimental results showed that the proposed model outperformed other models, such as the single model and random subspace ensemble model.

• Journal of Intelligence and Information Systems
• /
• v.23 no.1
• /
• pp.127-141
• /
• 2017

Advances in Internet technologies and the proliferation of mobile devices enabled consumers to approach a wide range of goods and services, while causing an adverse effect that they have hard time reaching their congenial items even if they devote much time to searching for them. Accordingly, businesses are using the recommender systems to provide tools for consumers to find the desired items more easily. Association Rule Mining (ARM) technology is advantageous to recommender systems in that ARM provides intuitive form of a rule with interestingness measures (support, confidence, and lift) describing the relationship between items. Given an item, its relevant items can be distinguished with the help of the measures that show the strength of relationship between items. Based on the strength, the most pertinent items can be chosen among other items and exposed to a given item's web page. However, the diversity of the measures may confuse which items are more recommendable. Given two rules, for example, one rule's support and confidence may not be concurrently superior to the other rule's. Such discrepancy of the measures in distinguishing one rule's superiority from other rules may cause difficulty in selecting proper items for recommendation. In addition, in an online environment where a web page or mobile screen can provide a limited number of recommendations that attract consumer interest, the prudent selection of items to be included in the list of recommendations is very important. The exposure of items of little interest may lead consumers to ignore the recommendations. Then, such consumers will possibly not pay attention to other forms of marketing activities. Therefore, the measures should be aligned with the probability of consumer's acceptance of recommendations. For this reason, this study proposes a model-based approach to combine those measures into one unified measure that can consistently determine the ranking of recommended items. A regression model was designed to describe how well the measures (independent variables; i.e., support, confidence, and lift) explain consumer's acceptance of recommendations (dependent variables, hit rate of recommended items). The model is intuitive to understand and easy to use in that the equation consists of the commonly used measures for ARM and can be used in the estimation of hit rates. The experiment using transaction data from one of the Korea's largest online shopping malls was conducted to show that the proposed model can improve the hit rates of recommendations. From the top of the list to 13th place, recommended items in the higher rakings from the proposed model show the higher hit rates than those from the competitive model's. The result shows that the proposed model's performance is superior to the competitive model's in online recommendation environment. In a web page, consumers are provided around ten recommendations with which the proposed model outperforms. Moreover, a mobile device cannot expose many items simultaneously due to its limited screen size. Therefore, the result shows that the newly devised recommendation technique is suitable for the mobile recommender systems. While this study has been conducted to cover the cross-selling in online shopping malls that handle merchandise, the proposed method can be expected to be applied in various situations under which association rules apply. For example, this model can be applied to medical diagnostic systems that predict candidate diseases from a patient's symptoms. To increase the efficiency of the model, additional variables will need to be considered for the elaboration of the model in future studies. For example, price can be a good candidate for an explanatory variable because it has a major impact on consumer purchase decisions. If the prices of recommended items are much higher than the items in which a consumer is interested, the consumer may hesitate to accept the recommendations.

• Journal of Intelligence and Information Systems
• /
• v.21 no.4
• /
• pp.93-110
• /
• 2015

The recommender system is a system which recommends products to the customers who are likely to be interested in. Based on automated information filtering technology, various recommender systems have been developed. Collaborative filtering (CF), one of the most successful recommendation algorithms, has been applied in a number of different domains such as recommending Web pages, books, movies, music and products. But, it has been known that CF has a critical shortcoming. CF finds neighbors whose preferences are like those of the target customer and recommends products those customers have most liked. Thus, CF works properly only when there's a sufficient number of ratings on common product from customers. When there's a shortage of customer ratings, CF makes the formation of a neighborhood inaccurate, thereby resulting in poor recommendations. To improve the performance of CF based recommender systems, most of the related studies have been focused on the development of novel algorithms under the assumption of using a single profile, which is created from user's rating information for items, purchase transactions, or Web access logs. With the advent of big data, companies got to collect more data and to use a variety of information with big size. So, many companies recognize it very importantly to utilize big data because it makes companies to improve their competitiveness and to create new value. In particular, on the rise is the issue of utilizing personal big data in the recommender system. It is why personal big data facilitate more accurate identification of the preferences or behaviors of users. The proposed recommendation methodology is as follows: First, multimodal user profiles are created from personal big data in order to grasp the preferences and behavior of users from various viewpoints. We derive five user profiles based on the personal information such as rating, site preference, demographic, Internet usage, and topic in text. Next, the similarity between users is calculated based on the profiles and then neighbors of users are found from the results. One of three ensemble approaches is applied to calculate the similarity. Each ensemble approach uses the similarity of combined profile, the average similarity of each profile, and the weighted average similarity of each profile, respectively. Finally, the products that people among the neighborhood prefer most to are recommended to the target users. For the experiments, we used the demographic data and a very large volume of Web log transaction for 5,000 panel users of a company that is specialized to analyzing ranks of Web sites. R and SAS E-miner was used to implement the proposed recommender system and to conduct the topic analysis using the keyword search, respectively. To evaluate the recommendation performance, we used 60% of data for training and 40% of data for test. The 5-fold cross validation was also conducted to enhance the reliability of our experiments. A widely used combination metric called F1 metric that gives equal weight to both recall and precision was employed for our evaluation. As the results of evaluation, the proposed methodology achieved the significant improvement over the single profile based CF algorithm. In particular, the ensemble approach using weighted average similarity shows the highest performance. That is, the rate of improvement in F1 is 16.9 percent for the ensemble approach using weighted average similarity and 8.1 percent for the ensemble approach using average similarity of each profile. From these results, we conclude that the multimodal profile ensemble approach is a viable solution to the problems encountered when there's a shortage of customer ratings. This study has significance in suggesting what kind of information could we use to create profile in the environment of big data and how could we combine and utilize them effectively. However, our methodology should be further studied to consider for its real-world application. We need to compare the differences in recommendation accuracy by applying the proposed method to different recommendation algorithms and then to identify which combination of them would show the best performance.

• The Korean Journal of Financial Management
• /
• v.24 no.4
• /
• pp.75-106
• /
• 2007

In this study, it is analyzed whether oil price plays a major role in the pricing return on Koran stock market and examined why the covariance risk between oil and return on stock is different in each industry. Firstly, this study explores whether the expected rate of return on stock is pricing due to global oil price factors as a function of risk premium by using a two-factor APT. Also, it is examined whether spill-over effects of oil price volatility affect the beta risk to oil price. Considering the asymmetry of oil price volatility, we use the GJR model. As a result, it shows that oil price is an independent pricing factor and oil price volatility transmits to stock return in only electricity and electrical equipment. Secondly, the two step-analyzing process is introduced to find why the covariance between oil price factor and stock return is different in each industry. The first step is to study whether beta risk exists in each industry by using two proxy variables like size and liquidity as control variables. The second step is to grasp the systematic relationship between the difference of liquidity and size and beta to oil price factor by using the panel-data model which can be analyzed efficiently using the cross-sectional data formed with time series. Through the analysis, we can argue that oil price factor is an independent pricing factor in only electricity and electrical equipment having the greatest market capitalization, and know that beta risk to oil price factor is a proxy of size in the other industries. According to the result of panel-data model, it is argued that the beta to oil price factor augments when market capitalization increases and this fact supports the first assertion. In conclusion, the expected rate of return of electricity and electrical equipment works as a function of risk premium to market portfolio and oil price, and the reason to make beta risk power differentiated in each industry attributes to the size.

• 한국벤처창업학회:학술대회논문집
• /
• 2006.04a
• /
• pp.167-201
• /
• 2006

Korean corporations are transforming their vertical operational structure to a team-based structure to compete in the rapidly changing environment and for improved performance. However, a high percentage of the respondents in KBS said that despite the appearance of the present team structure, the organization operates much like a vertically-structured organization. This result can be attributed to the lack of study and implementation toward the goal of empowerment, the key variable for the success of the team-based structure. This study aims to provide policy suggestions on how to implement the process of empowerment, by investigating the conditions that hinder the process and the attitude of the KBS employees. For the cross-sectional study, this thesis examined the domestic and international references, conducted a survey of KBS employees, personal interviews and made direct observations. Approximately 1,200 copies of the Questionnaire were distributed and 474 were completed and returned. The analysis used SPSS 12.0 software to process the data collected from 460 respondents. For the longitudinal-study, six categories that were common to this study and "The Report of the Findings of KBS Employees' View of the Team Structure" were selected. The comparative study analyzed the changes in a ten-month period. The survey findings showed a decrease of 24.2%p in the number of responses expressing negative views of the team structure and a decrease of 1.29%p in the number of positive responses. The findings indicated a positive transformation illustrating employees' improved understanding and approval of the team structure. However, KBS must address the issue on an ongoing basis. It has been proven that the employee empowerment increases the productivity of the individual and the group. In order to boost the level of empowerment, the management must exercise new, innovative leadership and build trust between the managers and the employees first. Additional workload as a result of shirking at work places was prevalent throughout all divisions and ranks, according to the survey data. This outcome leads to the conclusion that the workload is not evenly distributed or shared. And the data also showed the employees do not trust the assessment and rewards system. More attention and consideration must be paid to the team size and job allocation in order to address this matter; the present assessment and rewards system need to be complemented. The type of leadership varies depending on the characteristics of the organization's structure and employees' disposition. KBS must develop and reform its own management, leadership style to suit the characteristics of individual teams. Finally, for a soft-landing of KBS team structure, in-house training and education are necessary.

• Journal of the Korean Society of Clothing and Textiles
• /
• v.30 no.12 s.159
• /
• pp.1690-1696
• /
• 2006

The purpose of this study is to provide the basic data useful in designing apparel patterns for aged abdominal obese women. The body measurements of 318 women were taken at random, whose ages were over 60 and fields of action were colleges, sports centers, or business sites in Seoul and the neighboring districts. A total of 33 features in the upper body and lower body were used fer the anthropometric measurement and analysis using anthropometry. The collected measurement data were processed statistically using the SPSS 12.0 program for technical statistical analysis, t-test, frequency analysis, correlation analysis. The results of the study are as follows. 1. Subjects were classified into two groups as a result of analysis for measurement data. It was revealed that 251(about 79 percent) women of total subjects(n=318) have a characteristic of abdominal obese body type and elderly women of these group usually had big abdomen rather than hip. The criteria of abdominal obesity based on waist-hip ratio, WHR(=0.85). 2. Aged abdominal obese women have shown much larger size in most body measurements except items of some vertical length, such as bust ponit-bust point, font interscye, back interscye with circumference and depth of armscye, bust, waist, abdomen and hip while showing no difference in height, biacrominal breadth, hip width, neck shoulder point to breast point, crotch length. 3. Vervaeck index(=100.1) and Rohrer index(=1.7) indicated that the abdominal obese women were fat in overall body. And aspect ratio of waist(=0.86), abdomen(=0.92) and hip(=0.75) also appeared high that the shape of cross sections in those regions was similar to a figure of circle 4. In view of the correlation coefficient between hip circumference and the rest measurement items, and between hip circumference inclusively of the abdomen protrusion and the rest measurement items, there were found some differences for each group. In case of Group (abdominal obese group), the former is smaller than the other. 5. In case of Abdominal obese women, hip circumference inclusively of the abdomen protrusion is more mutually related to the rest items related to make apparel pattern as waist circumference, depth of armscye and so on than what hip circumference is. This result indicated which must be considered hip circumference inclusively of the abdomen protrusion to make apparel patterns for abdominal obese women unlike women of common body types.

• Archives of design research
• /
• v.18 no.1 s.59
• /
• pp.167-174
• /
• 2005

This study is to objectively support the emotional and intuitional decision making of the designer by means of developing the supporting models and tools of color coordination. Based on the color grouping system and representative vocabularies suggested in the precedent 'Study on the Grouping System of Fabric Color,' this study suggested the manufacture of the supporting model of color coordination that could be used practically through the design of coloring group. The results of this study can be summarized as below. Firstly, 687 colors in total have been collected from the four world famous collections, the street fashion of 2002 F/W 2003 S/S Season and the representative brands in each group for five years from 1999 to 2003 in order to single out the basic colors for the purpose of composing the color groups. Secondly, 687 collected colors have been grouped into 144 colors in total through the three-step process for the extraction of coloring groups. Thirdly, the final extracted colors have been divided into , , , group by the grouping system specified in the precedent study and the said four large groups have been again subdivided into 12 small groups. Fourthly, the suggested colors in each group have established a color coordination system by introducing the concept of the crossover coordination that could be matched with other groups as well as the coordination within the group. Fifthly, we have dyed 144 colors in total that have consisted of the coloring system of four representative groups (twelve subgroups) in each methodical tone as in the above in cotton yarn, one of the representative materials in fabric fashion design industry. Besides, we have specified the symbol of the Pantone Color Book and CMYK values in each color that has consisted of the system considering the industrial characteristics of fashion as a global business and the compatibility with the related design industry. Sixthly, we have packed the completed yam made of fabrics in the designed container for the easy use of cross-coordination and have completed a color coordination system that could be easily utilized for the fashion-related working-level staffs.

• Journal of Distribution Science
• /
• v.9 no.4
• /
• pp.43-50
• /
• 2011

Concerning the risk factors of the outbreak of a fire in a traditional market, most of those markets are located in downtown areas or residential areas; thus, although their location may be favorable in terms of marketability, they face a potential risk in that a fire may develop into a large blaze owing to poor environment or the absence of facilities prepared for disaster during a fire. Moreover, as many people are densely poised in the markets, it is very probable that a fire may occur owing to the excessive use of heaters in the winter as well as the reckless use of electric and gas facilities. It seems that traditional markets encounter difficulty being insured against fire, because of their vulnerability and that the vast majority of small-scale sellers are likely to suffer mental anguish and tremendous physical injury in case of a fire. However, most of those sellers in the traditional markets are hand-to-mouth sellers, and they lack awareness of safety concerns and have insufficient experience in safe facility management. As small-scale sellers constitute the majority in the traditional market, the subscription rate of fire insurance in most of the traditional markets is low for the reasons of their needy circumstances and their financial burden. Statistically, the subscription by street vendors is non-existent; therefore, these vendors have a fairly limited access to indemnification after fire damage. Because of these problems, this study's purpose is to identify the current level of insurance subscription by these markets, which are exposed to poor facilities and vulnerability to fire. In order to fix this, it appears that shop owners and consumers will have to band together. For this study, we executed a fire policyholder fact-finding mission at traditional markets with approximately 108 and 981 stores. The research method was executed by an investigation using one-on-one individual interviews using a questionnaire. The contents investigated current insurance subscriptions. The method of analysis looked at the difference of insured amount according to volume size through cross-tabulation of the difference of insured amount by possession form, difference of insured amount by market form, difference of insured amount by category of business, difference of insured amount by market size, etc. Furthermore, the study should be used to propose solutions for problems through theoretical review with the use of a literature research, because the field case study was through interviews with the persons concerned, and the survey of the current insurance subscriptions by traditional market shopkeepers. The traditional market would generally have difficulty affording fire insurance. Fire insurance subscription rates of most of the market proved to be inactive, because of the economic burden of payment. Lack of funds is thought to be the main factor that causes a lack of realization about the necessity of fire insurance. In addition to expensive insurance premiums, sometimes, the companies' valuation of the businesses is lower than their actual valuations, and they do not pay out enough during a claim. The research presents an improvement plan that, when presented at the traditional markets, may strengthen their ability to procure fire insurance through the help of the central government. Researchers connected with the traditional market mainly accomplish the initial research. However, although this research has its limitations, it offers considerable benefits. For future researchers, I would suggest looking at several regions for comparison.

• Journal of Intelligence and Information Systems
• /
• v.17 no.2
• /
• pp.61-75
• /
• 2011

A Geographic Information System(GIS) is a system that captures, stores, analyzes, manages and presents data with reference to geographic location data. In the later 1990s and earlier 2000s it was limitedly used in government sectors such as public utility management, urban planning, landscape architecture, and environmental contamination control. However, a growing number of open-source packages running on a range of operating systems enabled many private enterprises to explore the concept of viewing GIS-based sales and customer data over their own computer monitors. K telecommunication company has dominated the Korean telecommunication market by providing diverse services, such as high-speed internet, PSTN(Public Switched Telephone Network), VOLP (Voice Over Internet Protocol), and IPTV(Internet Protocol Television). Even though the telecommunication market in Korea is huge, the competition between major services providers is growing more fierce than ever before. Service providers struggled to acquire as many new customers as possible, attempted to cross sell more products to their regular customers, and made more efforts on retaining the best customers by offering unprecedented benefits. Most service providers including K telecommunication company tried to adopt the concept of customer relationship management(CRM), and analyze customer's demographic and transactional data statistically in order to understand their customer's behavior. However, managing customer information has still remained at the basic level, and the quality and the quantity of customer data were not enough not only to understand the customers but also to design a strategy for marketing and sales. For example, the currently used 3,074 legal regional divisions, which are originally defined by the government, were too broad to calculate sub-regional customer's service subscription and cancellation ratio. Additional external data such as house size, house price, and household demographics are also needed to measure sales potential. Furthermore, making tables and reports were time consuming and they were insufficient to make a clear judgment about the market situation. In 2009, this company needed a dramatic shift in the way marketing and sales activities, and finally developed a dedicated GIS_based market analysis and sales management system. This system made huge improvement in the efficiency with which the company was able to manage and organize all customer and sales related information, and access to those information easily and visually. After the GIS information system was developed, and applied to marketing and sales activities at the corporate level, the company was reported to increase sales and market share substantially. This was due to the fact that by analyzing past market and sales initiatives, creating sales potential, and targeting key markets, the system could make suggestions and enable the company to focus its resources on the demographics most likely to respond to the promotion. This paper reviews subjective and unclear marketing and sales activities that K telecommunication company operated, and introduces the whole process of developing the GIS information system. The process consists of the following 5 modules : (1) Customer profile cleansing and standardization, (2) Internal/External DB enrichment, (3) Segmentation of 3,074 legal regions into 46,590 sub_regions called blocks, (4) GIS data mart design, and (5) GIS system construction. The objective of this case study is to emphasize the need of GIS system and how it works in the private enterprises by reviewing the development process of the K company's market analysis and sales management system. We hope that this paper suggest valuable guideline to companies that consider introducing or constructing a GIS information system.

• Journal of Food Hygiene and Safety
• /
• v.31 no.4
• /
• pp.258-263
• /
• 2016

The risks of sanitary indicative bacteria, heavy metals and chlorinated derivatives in 94 cases of sanitary wet towels used in food services (39 from sanitary wet towel treatment business, 55 from food services) were assessed in the present study. Lead was detected in the range of N.D.~0.41 mg/kg (75 cases were not detected), N.D.~0.25 mg/kg of arsenic (93 cases were not detected), N.D.~0.01 mg/kg of cadmium (7 cases were lower than limit of quantitation; 87 cases were not detected), 0.003 mg/kg ~ 0.09 mg/kg of mercury. And chromium (VI) was not detected from all samples. The level of lead was the highest among the tested heavy metals, and the highest concentration of lead was 0.41 mg/kg. However, it was only 2.1% of legal limit (less than 20 mg/kg). The average moisture content of the samples was 61.9% (50.0% ~ 77.0%) and it showed no relevance to the detection of bacterial counts. Escherichia coli was not detected. Bacterial counts were detected 43 cases and among them, 24 cases were exceeded the legal limit. It was verified that the packaging conditions of sanitary wet towel (whether it is packed by a piece or not and sealed or not) are critical factors to cause the germ contamination and cross contamination in the wet towels. The chlorinated derivatives (chlorites and chlorates) were detected in 17 (19.3%) out of 88 cases. The results would be used as preliminary information to establish the programs of "Safety education for manufacturers and public policy of safety".