• Title/Summary/Keyword: performance test

Search Result 21,626, Processing Time 0.052 seconds

Developments of Local Festival Mobile Application and Data Analysis System Applying Beacon (비콘을 활용한 위치기반 지역축제 모바일 애플리케이션과 데이터 분석 시스템 개발)

  • Kim, Song I;Kim, Won Pyo;Jeong, Chul
    • Korea Science and Art Forum
    • /
    • v.31
    • /
    • pp.21-32
    • /
    • 2017
  • Local festivals form the regional cultures and atmosphere of communication; they increase the demand of domestic tourism businesses and thus, have an important role in ripple effects (e.g. regional image improvement, tourist influx, job creation, regional contents development, and local product sales) and economic revitalization. IoT (Internet of Thing) technologies have been developed especially, beacon-one of the IoT services has been applied as plenty of types and forms both domestically and internationally. However, notwithstanding expansion of current digital mobile technologies, it still remains as difficult for the individual to track the information about all the local festivals and to fulfill the tourists' needs of enjoying festivals given the weak strategic approaches and advertisement activities. Furthermore, current festival-related mobile applications don't function well as delivering information and have numerous contents issues (e.g. ways of information delivery within the festival places, independent application usage for each festival, one time usage due to one time event). This research, based on the background mentioned above, aims to develop the local festival mobile application and data analysis system applying beacon technology. First of all, three algorithms were developed, namely, 'festival crowding algorithm', 'visitor stats algorithm', and 'customized information algorithm', and then beta test was followed with the developed application and data analysis system. As a result, they could form the database of visitors' types and behaviors, and provide functions and services, such as personalized information, waiting time for festival contents, and 'hot place' function. Besides, in Google Play store, they also got the titles given with more than 13,000 downloads within first three months and as the most exposed application related with festivals; and, thus, got credited with their marketability and excellence. This research follows this order: chapter 2 shows the literature review of local festival related with technology development, beacon service, and festival application. In Chapter 3, design plans and conditions are described of developing local festival mobile application and data analysis system with beacon. Chapter 4 evaluates the results of the beta performance test to verify applicability of the developed application and data analysis system, and lastly, chapter 5 explains the conclusion and suggests the future research.

Research on Radiation Shielding Film for Replacement of Lead(Pb) through Roll-to-Roll Sputtering Deposition (롤투롤 스퍼터링 증착을 통한 납(Pb) 대체용 방사선 차폐필름 개발)

  • Sung-Hun Kim;Jung-Sup Byun;Young-Bin Ji
    • Journal of the Korean Society of Radiology
    • /
    • v.17 no.3
    • /
    • pp.441-447
    • /
    • 2023
  • Lead(Pb), which is currently mainly used for shielding purposes in the medical radiation, has excellent radiation shielding functions, but is continuously exposed to radiation directly or indirectly due to the harmfulness of lead itself to the human body and the inconvenience caused by its heavy weight. Research on shielding materials that are human-friendly, lightweight, and convenient to use that can block risks and replace lead is continuously being conducted. In this study, based on the commonly used polyethylene terephthalate (PET) film and the fabric material used in actual radiation protective clothing, a multi-layer thin film was realized through sputtering and vacuum deposition of bismuth, tungsten, and tin, which are metal materials that can shield radiation. Thus, a shielding film was produced and its applicability as a radiation shielding material was evaluated. The radiation shielding film was manufactured by establishing the optimized conditions for each shielding material while controlling the applied voltage, roll driving speed, and gas supply amount to manufacture the shielding film. The adhesion between the parent material and the shielding metal thin film was confirmed by Cross-cut 100/100, and the stability of the thin film was confirmed through a hot water test for 1 hour to measure the change of the thin film over time. The shielding performance of the finally realized shielding film was measured by the Korea association for radiation application (KARA), and the test conditions (inverse wide beam, tube voltage 50 kV, half layer 1.828 mmAl) were set to obtain an attenuation ratio of 16.4 (initial value 0.300 mGy/s, measured value 0.018 mGy/s) and damping ratio 4.31 (initial value 0.300 mGy/s, measured value 0.069 mGy/s) were obtained. by securing process efficiency for future commercialization, light and shielding films and fabrics were used to lay the foundation for the application of films to radiation protective clothing or construction materials with shielding functions.

Verification of Multi-point Displacement Response Measurement Algorithm Using Image Processing Technique (영상처리기법을 이용한 다중 변위응답 측정 알고리즘의 검증)

  • Kim, Sung-Wan;Kim, Nam-Sik
    • KSCE Journal of Civil and Environmental Engineering Research
    • /
    • v.30 no.3A
    • /
    • pp.297-307
    • /
    • 2010
  • Recently, maintenance engineering and technology for civil and building structures have begun to draw big attention and actually the number of structures that need to be evaluate on structural safety due to deterioration and performance degradation of structures are rapidly increasing. When stiffness is decreased because of deterioration of structures and member cracks, dynamic characteristics of structures would be changed. And it is important that the damaged areas and extent of the damage are correctly evaluated by analyzing dynamic characteristics from the actual behavior of a structure. In general, typical measurement instruments used for structure monitoring are dynamic instruments. Existing dynamic instruments are not easy to obtain reliable data when the cable connecting measurement sensors and device is long, and have uneconomical for 1 to 1 connection process between each sensor and instrument. Therefore, a method without attaching sensors to measure vibration at a long range is required. The representative applicable non-contact methods to measure the vibration of structures are laser doppler effect, a method using GPS, and image processing technique. The method using laser doppler effect shows relatively high accuracy but uneconomical while the method using GPS requires expensive equipment, and has its signal's own error and limited speed of sampling rate. But the method using image signal is simple and economical, and is proper to get vibration of inaccessible structures and dynamic characteristics. Image signals of camera instead of sensors had been recently used by many researchers. But the existing method, which records a point of a target attached on a structure and then measures vibration using image processing technique, could have relatively the limited objects of measurement. Therefore, this study conducted shaking table test and field load test to verify the validity of the method that can measure multi-point displacement responses of structures using image processing technique.

Responses of Health Physical Fitness and Arterial Stiffness through Cigarette Smoking (흡연습관이 성인 남성의 건강관련체력 및 동맥경직도에 미치는 영향)

  • Jung, Min-Kyung;Park, Eun-Kyung;Yoo, Jae-Hyun
    • Journal of Korea Entertainment Industry Association
    • /
    • v.13 no.2
    • /
    • pp.197-205
    • /
    • 2019
  • This study was to compare arterial stiffness and hemodynamic responses between male smokers and non-smokers. This study also investigates the influences of smoking before exercise on arterial stiffness and hemodynamic responses. 24 male subjects of age 20-29 without history of cardiorespiratory disease were divided into smokers and non-smokers. Smokers had more than 5 years of smoking experience. In order to evaluate the effects of pre-exercise smoking, smokers were tested twice, once with a cigarette before the exercise and the other once without one. Data was collected from bio-impedance analysis, SphygmoCor XCEL, graded exercise test, and fitness test. Main results of this study are as follows: First, there are differences between smokers and non-smokers in cardiorespiratory and hemodynamic response functions, as shown by maximal oxygen consumption, exercise duration, and heart rate. Second, the although the arterial stiffness between smokers and non-smokers showed statistically significant differences in the speed of the pulse wave velocity and augmentation index, smoker had a faster rate. It shows that smoking behavior has a negative impact on the cardiovascular system. Third, pre-exercise smoking behavior does have an impact on cardiorespiratory and hemodynamic response functions, as shown by exercise duration and heart rate. Lastly, arterial stiffness between smokers and non-smokers showed statistically not significant in the speed of the pulse wave velocity and augmentation index. However, the difference was not statistically significant. Brachial systolic pressure, a component of pulse wave analysis, on the other hand, was significantly dependent on pre-exercise smoking behavior. Subjects who participated in this study are college students in early 20s. Given their relatively short history of smoking, it is possible that their smoking habits are not severe enough to develop into cardiorespiratory or cardiovascular diseases. But Smokers showed lower levels of cardiopulmonary functions, as maximal oxygen consumption and exercise duration than nonsmokers.

A study of the Medical System in the Early Chosun-Dynasty (조선시대(朝鮮時代) 전기(前期)의 의료제도(醫療制度)에 대한 연구(硏究))

  • Han, Dae-Hee;Kang, Hyo-Shin
    • Journal of Korean Medical classics
    • /
    • v.9
    • /
    • pp.555-652
    • /
    • 1996
  • Up to the present the scholastic achievements in the history of the medical system have been rather scare despite its importance in the Korean History. Hence, this dissertation attempts to examine the significance of the institute in the Korean History, covering the period from the ancient times through the early Chosun-Dynasty. In the ancient times, the medical practice relied primarily upon human instincts and experiences at the same time, shaman's incantations were widely believed to cure diseases, the workings of evil spirits supposedly. For the period from the Old Chosun through Samhan(巫堂), Chinese refugees brought a long medical knowledge and skills of the continent. New Chinese medicine, traditional practices and incantations were generally used at this time. Medicine and the medical system were arranged by the period of the Three Countries(三國時代). No definite record concerning Koguryo remains now. As for Paekje, however, history shows that they set up the system under the Chinese influence, assigning medical posts such as Euibaksa(medical doctor), Chaeyaksa(pharmacist), and Jukeumsa(medicine man) within Yakbu(department of medicine). Scientifically advanced, they sent experts to Japan, giving a tremendous influence on the development of the science on ancient Japan. After the unification of the three countries, Shilla had their own system after the model of Dang(唐). This system of the Unified Shilla was continued down to Koryo and became the backbone of the future ones. In the ancient time religion and medicine were closely related. The curative function of the shaman was absolute. Buddhism played a notable part in medical practice, too, producing numerous medical monks. The medical system of Koryo followed the model of Dang with some borrowings from Song(宋). Sangyakkuk(尙藥局) was to deal exclusively with the diseases of the monarch whereas Taeeuigam(太醫監) was the central office to handle the national medical administration and the qualification test and education for doctors. In addition, Dongsodaebiwon(東西大悲院), Jewibo(濟危寶), and Hyeminkuk(惠民局) were public hospitals for the people, and a few aristocrats practiced medicine privately. In 987, the 6th year of Songjong(成宗), local medical operations were installed for curing the sick and educating medical students. Later Hyonjong(顯宗), established Yakjom(clinics, 藥店) throughout the country and officials were sent there to see patients. Foreign experts, mainly from Song, were invited frequently to deliver their advanced technology, and contributed to the great progress of the science in Korea. Medical officials were equipped with better land and salary than others, enjoying appropriate social respect. Koryo exchanged doctors, medicine and books mainly with Song, but also had substantial interrelations with Yuan(元), Ming(明), Kitan(契丹), Yojin(女眞), and Japan. Among them, however, Song was most influential to the development of medicine in Koryo. During Koryo Dynasty Buddhism, the national religion at the time, exercised bigger effect on medicine than in any other period. By conducting national ceremonies and public rituals to cure diseases, Taoism also affected the way people regarded illness. Curative shamanism was still in practice as well. These religious practices, however, were now engaged only when medication was already in use or when medicine could not held not help any more. The advanced medical system of Koryo were handed down to Chosun and served the basis for further progress. Hence, then played well the role to connect the ancient medicine and the modern one. The early Chosun followed and systemized the scientific and technical achievement in medicine during the Koryo Dynasty, and furthermore, founded the basis of the future developments. Especially the 70 years approximately from the reign of Sejong(世宗) to that of Songjong(成宗) withnessed a termendous progress in the field with the reestablishment of the medical system. The functions of the three medical institute Naeeuiwon(內醫院), Joneuigam(典醫監), Hyeminkuk(惠民局) were expanded. The second, particualy, not only systemized all the medical practices of the whole nation, but also grew and distributed domestic medicaments which had been continually developed since the late Koryo period. In addition, Hyeminso(惠民局, Hwarinwon(活人院)) and Jesaenwon(濟生院)(later merged to the first) played certain parts in the curing illness. Despite the active medical education in the capital and the country, the results were not substantial, for the aristocracy avoided the profession due to the social prejudice against technicians including medical docotors. During the early Chosun-Dynasty, the science was divided into Chimgueui (acupuncturist), Naryogeui(specialist in scrofula) and Chijongeui (specialist in boil). For the textbooks, those for the qualification exam were used, including several written by the natives. With the introduction on Neoconfucianism(性理學) which reinforced sexual segregation, female doctors appeared for the female patients who refused to be seen by male doctors. This system first appeared in 1406, the sixth year of Taejong(太宗), but finally set up during the reign of Sejong. As slaves to the offices, the lowest class, female doctors drew no respect. However, this is still significant in the aspect of women's participation in society. They were precedents of midwives. Medical officials were selected through the civil exam and a special test. Those who passed exams were given temporary jobs and took permanent posts later. At that time the test score, the work experience and the performance record of the prospective doctor were all taken into consideration, for it was a specialized office. Most doctors were given posts that changed every six months, and therefore had fewer chances for a goverment office than the aristocracy. At the beginning the social status of those in medicine was not that low, but with the prejudice gradully rising among the aristocracy, it became generally agreed to belong to the upper-middle technician class. Dealing with life, however, they received social respect and courtesy from the public. Sometimes they collected wealth with their skills. They kept improving techniques and finally came to take an important share in modernization process during the late Chosun-Dynasty.

  • PDF

Machine learning-based corporate default risk prediction model verification and policy recommendation: Focusing on improvement through stacking ensemble model (머신러닝 기반 기업부도위험 예측모델 검증 및 정책적 제언: 스태킹 앙상블 모델을 통한 개선을 중심으로)

  • Eom, Haneul;Kim, Jaeseong;Choi, Sangok
    • 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.

Studies on the Efficacy of the Newly Developed Automatic Liquid Feeder for 7-day Old Early Weaned Piglets (7일령 조기 이유자돈에 대한 액상사료 자동급이기 이용효과)

  • Yoo, Y.H.;Chung, I.B.;Jang, B.G.;Moon, H.K.;Kim, T.I.;Han, J.D.;Park, H.S.
    • Journal of Animal Science and Technology
    • /
    • v.45 no.6
    • /
    • pp.1079-1088
    • /
    • 2003
  • This study was conducted to test the efficacy of the automatic liquid feeder(designated as NLRI) developed by National Livestock Research Institute, Korea for 7-day old early-weaning piglets. The other automatic liquid feeder imported from USA(designated as IALF) was used for a comparative purpose. A total of sixty piglets of 7 days of age were allotted to three treatments. The control group of 20 piglets was raised by their own sows until 21 days of age, while others were shifted to automatic liquid feeders of either NLRI or IALF on 7 days of age and reared during 14 days post-weaning. After then, all the piglets were fed nutritionally identical mash diets until 70 days of age. Compared with control group, body weight of piglets reared on automatic liquid feeders were lighter(P〈0.05) at 21 days of age [5.55kg (control) vs 4.97 (NLRI) and 4.98kg (IALF)], while heavier(P〈0.05) at 70 days of age(24.82kg vs 30.17 and 29.42kg). The results indicated that pigs reared on liquid feeding showed higher(P〈0.05) average daily gain than control pigs during the whole experimental period [346.7g (control) vs 425.8 (NLRI) and 416.1g (IALF)], while no difference was found in feed/gain (1.67 vs 1.78 and 1.84). There was no difference in growth performance and incidence of diarrhea between two automatic liquid feeders, NLRI and IALF. Compared with control group, intestinal villi tended to be shorter in liquid feeding group during the first week, but were recovered within two weeks. The results suggest that the automatic liquid feeder newly developed by National Livestock Research Institute, Korea can be successfully used for rearing young piglets weaned at very early age.

Development of a Solar Collector Performance of Cylindrical Parabolic Concentrating Solar Collector (태양열(太陽熱) 집열기개발(集熱器開發)에 관(關)한 연구(硏究) - 포물반사곡면(抛物反射曲面)으로된 2차원(二次元) 집광식(集光式) 태양열(太陽熱) 집열기(集熱器)의 성능분석(性能分析) -)

  • Song, Hyun Kap;Yon, Kwang Seok;Cho, Sung Chan
    • Journal of Biosystems Engineering
    • /
    • v.10 no.1
    • /
    • pp.54-68
    • /
    • 1985
  • It is desirable to collect the solar thermal energy at relatively high temperature in order to minimize the size of thermal storage system and to enlarge the scope of solar thermal energy utilization. So far the concentrating solar collector has been developed to collect solar thermal energy at relatively high temperature, but it has some difficulties in maintaining the volumetric body of solar collector for long term utilization. On the other hand, the flat-plate solar collector has been developed to collect the solar thermal energy at low temperature, and it has advantages in maintaining the system for long term utilization, since it's thickness is thin and not volumetric. In this study, to develop a solar collector that has both advantages of collecting solar thermal energy at high temperature and fixing conveniently the collector system for long term period, a cylindrical parabolic concentrating solar collector was designed, which has two rows of parabolic reflectors and thin thickness such as the flat-plate solar collector, maintaining the optical form of concentrating solar collector. The characteristics of the concentrating parabolic solar collector newly designed was analysed and the results are summarized as follows; 1. The temperature of the air enclosed in solar collector was all the same as $50^{\circ}C$ in both cases of the open and closed loop, and when the heat transfer fluid was not circulated in tubular absorber, the maximum surface temperature of the absorber was $118-120^{\circ}C$, this results suggested that the heat transfer fluid could be heated up to $118^{\circ}C$. 2. In case of longitudinal installation of the solar collector, the temperature difference of heat transfer fluid between inlet and outlet was $4^{\circ}-6^{\circ}C$ at the flow rate of $110-130{\ell}/hr$, and the collected solar energy per unit area of collector was $300-465W/m^2$. 3. The collected solar energy per unit area for 7 hours was 1960 Kcal/$m^2$ for the open loop and 220 Kcal/$m^2$ for the closed loop. Therefore it is necessary to combine the open and closed loop of solar collectors to improve the thermal efficiency of solar collector. 4. The thermal efficiency of the solar collector (C.P.C.S.C.) was proportional to the density of solar radiation, indicating the maximum thermal efficiency ${\eta}_{max}=58%$ with longitudinal installation and ${\eta}_{max}=45%$ with lateral installation. 5. The thermal efficiency of the solar collector (C.P.C.S.C.) was increased in accordance with the increase of flow rate of heat transfer fluid, presenting the flow rate of $110{\ell}/hr$ was the value of turning point of the increasing rate of the collector efficiency, therefore the flow rate of $110{\ell}/hr$ was considered as optimum value for the test of the solar collector (C.P.C.S.C.) performance when the heat transfer fluid is a liquid. 6. In both cases of longitudinal and lateral installation of the solar collector (C.P.C.S.C.), the thermal efficiency was decreased linearly with an increase in the value of the term ($T_m-T_a$)/Ic and the increasing rate of the thermal efficiency was not effected by the installation method of solar collector.

  • PDF

Investigating Dynamic Mutation Process of Issues Using Unstructured Text Analysis (부도예측을 위한 KNN 앙상블 모형의 동시 최적화)

  • Min, Sung-Hwan
    • 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.

Enhancing Predictive Accuracy of Collaborative Filtering Algorithms using the Network Analysis of Trust Relationship among Users (사용자 간 신뢰관계 네트워크 분석을 활용한 협업 필터링 알고리즘의 예측 정확도 개선)

  • Choi, Seulbi;Kwahk, Kee-Young;Ahn, Hyunchul
    • Journal of Intelligence and Information Systems
    • /
    • v.22 no.3
    • /
    • pp.113-127
    • /
    • 2016
  • Among the techniques for recommendation, collaborative filtering (CF) is commonly recognized to be the most effective for implementing recommender systems. Until now, CF has been popularly studied and adopted in both academic and real-world applications. The basic idea of CF is to create recommendation results by finding correlations between users of a recommendation system. CF system compares users based on how similar they are, and recommend products to users by using other like-minded people's results of evaluation for each product. Thus, it is very important to compute evaluation similarities among users in CF because the recommendation quality depends on it. Typical CF uses user's explicit numeric ratings of items (i.e. quantitative information) when computing the similarities among users in CF. In other words, user's numeric ratings have been a sole source of user preference information in traditional CF. However, user ratings are unable to fully reflect user's actual preferences from time to time. According to several studies, users may more actively accommodate recommendation of reliable others when purchasing goods. Thus, trust relationship can be regarded as the informative source for identifying user's preference with accuracy. Under this background, we propose a new hybrid recommender system that fuses CF and social network analysis (SNA). The proposed system adopts the recommendation algorithm that additionally reflect the result analyzed by SNA. In detail, our proposed system is based on conventional memory-based CF, but it is designed to use both user's numeric ratings and trust relationship information between users when calculating user similarities. For this, our system creates and uses not only user-item rating matrix, but also user-to-user trust network. As the methods for calculating user similarity between users, we proposed two alternatives - one is algorithm calculating the degree of similarity between users by utilizing in-degree and out-degree centrality, which are the indices representing the central location in the social network. We named these approaches as 'Trust CF - All' and 'Trust CF - Conditional'. The other alternative is the algorithm reflecting a neighbor's score higher when a target user trusts the neighbor directly or indirectly. The direct or indirect trust relationship can be identified by searching trust network of users. In this study, we call this approach 'Trust CF - Search'. To validate the applicability of the proposed system, we used experimental data provided by LibRec that crawled from the entire FilmTrust website. It consists of ratings of movies and trust relationship network indicating who to trust between users. The experimental system was implemented using Microsoft Visual Basic for Applications (VBA) and UCINET 6. To examine the effectiveness of the proposed system, we compared the performance of our proposed method with one of conventional CF system. The performances of recommender system were evaluated by using average MAE (mean absolute error). The analysis results confirmed that in case of applying without conditions the in-degree centrality index of trusted network of users(i.e. Trust CF - All), the accuracy (MAE = 0.565134) was lower than conventional CF (MAE = 0.564966). And, in case of applying the in-degree centrality index only to the users with the out-degree centrality above a certain threshold value(i.e. Trust CF - Conditional), the proposed system improved the accuracy a little (MAE = 0.564909) compared to traditional CF. However, the algorithm searching based on the trusted network of users (i.e. Trust CF - Search) was found to show the best performance (MAE = 0.564846). And the result from paired samples t-test presented that Trust CF - Search outperformed conventional CF with 10% statistical significance level. Our study sheds a light on the application of user's trust relationship network information for facilitating electronic commerce by recommending proper items to users.