• Journal of Intelligence and Information Systems
• /
• v.27 no.2
• /
• pp.1-15
• /
• 2021

If it is difficult for us to make decisions, we ask for advice from friends or people around us. When we decide to buy products online, we read anonymous reviews and buy them. With the advent of the Data-driven era, IT technology's development is spilling out many data from individuals to objects. Companies or individuals have accumulated, processed, and analyzed such a large amount of data that they can now make decisions or execute directly using data that used to depend on experts. Nowadays, the recommender system plays a vital role in determining the user's preferences to purchase goods and uses a recommender system to induce clicks on web services (Facebook, Amazon, Netflix, Youtube). For example, Youtube's recommender system, which is used by 1 billion people worldwide every month, includes videos that users like, "like" and videos they watched. Recommended system research is deeply linked to practical business. Therefore, many researchers are interested in building better solutions. Recommender systems use the information obtained from their users to generate recommendations because the development of the provided recommender systems requires information on items that are likely to be preferred by the user. We began to trust patterns and rules derived from data rather than empirical intuition through the recommender systems. The capacity and development of data have led machine learning to develop deep learning. However, such recommender systems are not all solutions. Proceeding with the recommender systems, there should be no scarcity in all data and a sufficient amount. Also, it requires detailed information about the individual. The recommender systems work correctly when these conditions operate. The recommender systems become a complex problem for both consumers and sellers when the interaction log is insufficient. Because the seller's perspective needs to make recommendations at a personal level to the consumer and receive appropriate recommendations with reliable data from the consumer's perspective. In this paper, to improve the accuracy problem for "appropriate recommendation" to consumers, the recommender systems are proposed in combination with context-based deep learning. This research is to combine user-based data to create hybrid Recommender Systems. The hybrid approach developed is not a collaborative type of Recommender Systems, but a collaborative extension that integrates user data with deep learning. Customer review data were used for the data set. Consumers buy products in online shopping malls and then evaluate product reviews. Rating reviews are based on reviews from buyers who have already purchased, giving users confidence before purchasing the product. However, the recommendation system mainly uses scores or ratings rather than reviews to suggest items purchased by many users. In fact, consumer reviews include product opinions and user sentiment that will be spent on evaluation. By incorporating these parts into the study, this paper aims to improve the recommendation system. This study is an algorithm used when individuals have difficulty in selecting an item. Consumer reviews and record patterns made it possible to rely on recommendations appropriately. The algorithm implements a recommendation system through collaborative filtering. This study's predictive accuracy is measured by Root Mean Squared Error (RMSE) and Mean Absolute Error (MAE). Netflix is strategically using the referral system in its programs through competitions that reduce RMSE every year, making fair use of predictive accuracy. Research on hybrid recommender systems combining the NLP approach for personalization recommender systems, deep learning base, etc. has been increasing. Among NLP studies, sentiment analysis began to take shape in the mid-2000s as user review data increased. Sentiment analysis is a text classification task based on machine learning. The machine learning-based sentiment analysis has a disadvantage in that it is difficult to identify the review's information expression because it is challenging to consider the text's characteristics. In this study, we propose a deep learning recommender system that utilizes BERT's sentiment analysis by minimizing the disadvantages of machine learning. This study offers a deep learning recommender system that uses BERT's sentiment analysis by reducing the disadvantages of machine learning. The comparison model was performed through a recommender system based on Naive-CF(collaborative filtering), SVD(singular value decomposition)-CF, MF(matrix factorization)-CF, BPR-MF(Bayesian personalized ranking matrix factorization)-CF, LSTM, CNN-LSTM, GRU(Gated Recurrent Units). As a result of the experiment, the recommender system based on BERT was the best.

• Korean Journal of Remote Sensing
• /
• v.38 no.6_2
• /
• pp.1691-1707
• /
• 2022

Cheonji Lake, the caldera of Baekdu Mountain, located on the border of the Korean Peninsula and China, alternates between melting and freezing seasonally. There is a magma chamber beneath Cheonji, and variations in the magma chamber cause volcanic antecedents such as changes in the temperature and water pressure of hot spring water. Consequently, there is an abnormal region in Cheonji where ice melts quicker than in other areas, freezes late even during the freezing period, and has a high-temperature water surface. The abnormal area is a discharge region for hot spring water, and its ice gradient may be used to monitor volcanic activity. However, due to geographical, political and spatial issues, periodic observation of abnormal regions of Cheonji is limited. In this study, the degree of ice change in the optimal region was quantified using a Landsat -5/-7/-8 optical satellite image and a Modified U-Net regression model. From January 22, 1985 to December 8, 2020, the Visible and Near Infrared (VNIR) band of 83 Landsat images including anomalous regions was utilized. Using the relative spectral reflectance of water and ice in the VNIR band, unique data were generated for quantitative ice variability monitoring. To preserve as much information as possible from the visible and near-infrared bands, ice gradient was noticed by applying it to U-Net with two encoders, achieving good prediction accuracy with a Root Mean Square Error (RMSE) of 140 and a correlation value of 0.9968. Since the ice change value can be seen with high precision from Landsat images using Modified U-Net in the future may be utilized as one of the methods to monitor Baekdu Mountain's volcanic activity, and a more specific volcano monitoring system can be built.

• Journal of Korean Society of Forest Science
• /
• v.90 no.2
• /
• pp.197-209
• /
• 2001

TOPMODEL, semi-distributed hydrological model, is frequently applied to predict the amount of discharge, main flow pathways and water quality in a forested catchment, especially in a spatial dimension. TOPMODEL is a kind of conceptual model, not physical one. The main concept of TOPMODEL is constituted by the topographic index and soil transmissivity. Two components can be used for predicting the surface and subsurface contributing area. This study is conducted for the validation of applicability of TOPMODEL at small forested catchments in Korea. The experimental area is located at Gwangneung forest operated by Korea Forest Research Institute, Gyeonggi-do near Seoul metropolitan. Two study catchments in this area have been working since 1979 ; one is the natural mature deciduous forest(22.0 ha) about 80 years old and the other is the planted young coniferous forest(13.6 ha) about 22 years old. The data collected during the two events in July 1995 and June 2000 at the mature deciduous forest and the three events in July 1995 and 1999, August 2000 at the young coniferous forest were used as the observed data set, respectively. The topographic index was calculated using $10m{\times}10m$ resolution raster digital elevation map(DEM). The distribution of the topographic index ranged from 2.6 to 11.1 at the deciduous and 2.7 to 16.0 at the coniferous catchment. The result of the optimization using the forecasting efficiency as the objective function showed that the model parameter, m and the mean catchment value of surface saturated transmissivity, $lnT_0$ had a high sensitivity. The values of the optimized parameters for m and InT_0 were 0.034 and 0.038; 8.672 and 9.475 at the deciduous and 0.031, 0.032 and 0.033; 5.969, 7.129 and 7.575 at the coniferous catchment, respectively. The forecasting efficiencies resulted from the simulation using the optimized parameter were comparatively high ; 0.958 and 0.909 at the deciduous and 0.825, 0.922 and 0.961 at the coniferous catchment. The observed and simulated hyeto-hydrograph shoed that the time of lag to peak coincided well. Though the total runoff and peakflow of some events showed a discrepancy between the observed and simulated output, TOPMODEL could overall predict a hydrologic output at the estimation error less than 10 %. Therefore, TOPMODEL is useful tool for the prediction of runoff at an ungaged forested catchment in Korea.

• Radiation Oncology Journal
• /
• v.22 no.1
• /
• pp.55-63
• /
• 2004

Purpose : To evaluate the reflection of tumor motion according to the planning CT scan time. Material and Methods : A model of N-shape, which moved aiong the longitudinal axis during the ventilation caused by a mechanical ventilator, was produced. The model was scanned by planning CT, while setting the relative CT scan time (T: CT scan time/ventilatory period) to 0.33, 0.50, 0.67, 0.75, 1.00, 1.337, and 1.537. In addition, three patients with non-small cell lung cancer who received stereotactic radiosurgery In the Department of Radiation Oncology, Asan Medical Center from 03/19/2002 to 05/21/2002 were scanned. Slow (10 Premier, Picker, scan time 2.0 seconds per slice) and fast CT scans (Lightspeed, GE Medical Systems, with a scan time of 0.8 second per slice) were peformed for each patient. The magnitude of reflected movement of the N-shaped model was evaluated by measuring the transverse length, which reflected the movement of the declined bar of the model at each slice. For patients' scans, all CT data sets were registered using a stereotactic body frame scale with the gross tumor volumes delineated in one CT image set. The volume and three-dimensional diameter of the gross tumor volume were measured and analyzed between the slow and fast CT scans. Results : The reflection degree of longitudinal movement of the model increased in proportion to the relative CT scan times below 1.00 7, but remained constant above 1.00 T Assuming the mean value of scanned transverse lengths with CT scan time 1.00 T to be $100\%$, CT scans with scan times of 0.33, 0.50, 0.57, and 0.75 T missed the tumor motion by 30, 27, 20, and $7.0\%$ respectively, Slow (scan time 2.0 sec) and Fast (scan time 0.8 sec) CT scans of three patients with longitudinal movement of 3, 5, and 10 mm measured by fluoroscopy revealed the increases in the diameter along the longitudinal axis Increased by 6.3, 17, and $23\%$ in the slow CT scans. Conculsion : As the relative CT scan time increased, the reflection of the respiratory tumor movement on planning CT also Increased, but remained constant with relative CT scan times above 1.00 T When setting the planning CT scan time above one respiration period (>1.00 T), only the set-up margin is needed to delineate the planning target volume. Therefore, therapeutic ratio can be increased by reducing the radiation dose delivered to normal lung tissue.

• Journal of Korean Tunnelling and Underground Space Association
• /
• v.23 no.6
• /
• pp.469-484
• /
• 2021

Rock mass classification results have a great influence on construction schedule and budget as well as tunnel stability in tunnel design. A total of 3,526 tunnels have been constructed in Korea and the associated techniques in tunnel design and construction have been continuously developed, however, not many studies have been performed on how to assess rock mass quality and grade more accurately. Thus, numerous cases show big differences in the results according to inspectors' experience and judgement. Hence, this study aims to suggest a more reliable rock mass classification (RMR) model using machine learning algorithms, which is surging in availability, through the analyses based on various rock and rock mass information collected from boring investigations. For this, 11 learning parameters (depth, rock type, RQD, electrical resistivity, UCS, V_p, V_s, Young's modulus, unit weight, Poisson's ratio, RMR) from 13 local tunnel cases were selected, 337 learning data sets as well as 60 test data sets were prepared, and 6 machine learning algorithms (DT, SVM, ANN, PCA & ANN, RF, XGBoost) were tested for various hyperparameters for each algorithm. The results show that the mean absolute errors in RMR value from five algorithms except Decision Tree were less than 8 and a Support Vector Machine model is the best model. The applicability of the model, established through this study, was confirmed and this prediction model can be applied for more reliable rock mass classification when additional various data is continuously cumulated.

• Journal of the Korean Association of Oral and Maxillofacial Surgeons
• /
• v.36 no.5
• /
• pp.366-374
• /
• 2010

Introduction: This study evaluated the capability of silk fibroin (SF) and recombinant human bone morphogenetic protein-2 loaded SF (SF-BMP) as a bone defect replacement matrix when grafted in a calvarial bone defect of rats in vivo. Materials and Methods: A total 70 calvarial critical size defects (5.0 mm in diameter) made on 35 adult female Sprague-Dawley rats were used in this study. The defects were transplanted with (1) rhBMP-2 loaded silk fibroin graft (SF-BMP: 0.8+$10\;{\mu}g$), (2) Silk fibroin (SF: $10\;{\mu}g$), and (3) no graft material (Raw). The samples were evaluated with soft x-rays, alkaline phosphatase activity, calcium/phosphate quantification, histological and histomorphometric analysis at postoperative 4 and 8 weeks. Results: The SF-BMP group ($48.86{\pm}14.92%$) had a significantly higher mean percentage bone area than the SF group ($24.96{\pm}11.01%$) at postoperative 4 weeks.(P<0.05) In addition, the SF-BMP group ($40.01{\pm}12.43%$) had a higher % bone area at postoperative 8 weeks than the SF group ($33.26{\pm}5.15%$). The mean ratio of gray scale levels to the host bone showed that the SF-BMP group ($0.67{\pm}0.08$) had a higher mean ratio level than the SF group ($0.61{\pm}0.09$) at postoperative 8 weeks. These differences were not statistically significant.(P=0.168 and P=0.243, respectively) The ratio of the calcium and phosphate contents of the SF-BMP ($0.93{\pm}0.22$) group was lower than that of the SF ($1.90{\pm}1.42$) group at postoperative 4 weeks. However, the SF-BMP group ($0.75{\pm}0.31$) had a higher Ca/$PO_4$ ratio than the SF ($0.68{\pm}0.04$) at postoperative 8 weeks. These differences were not statistically significant.(P=0.126 and P=0.627, respectively) For the bone-specific alkaline phosphatase (ALP) activity, which is recognized as a reliable indicator of the osteoblast function, the SF-BMP ($23.71{\pm}8.60\;U/L$) groups had a significantly higher value than the SF group ($12.65{\pm}6.47\;U/L$) at postoperative 4 weeks.(P<0.05) At postoperative 8 weeks, the SF-BMP ($21.65{\pm}10.02\;U/L$) group had a lower bone-specific ALP activity than the SF group ($16.72{\pm}7.35\;U/L$). This difference was not statistically significant.(P=0.263) For the histological evaluation, the SF-BMP group revealed less inflammation, lower foreign body reactions and higher bone healing than the SF group at postoperative 4 and 8 weeks. The SF group revealed more foreign body reactions at postoperative 4 weeks. However, this immunogenic reaction decreased and the remnant of grafted material was observed at postoperative 8 weeks. For histomorphometric analysis, the SF-BMP group had a significantly longer bone length to total length ratio than those of the SF group at postoperative 4 and 8 weeks.(P<0.05) Conclusion: The rhBMP-2 loaded silk fibroin graft revealed fewer immunoreactions and inflammation as well as more new bone formation than the pure silk fibroin graft. Therefore, silk fibroin may be a candidate scaffold for tissue engineered bone regeneration.

• Journal of Aquaculture
• /
• v.10 no.3
• /
• pp.261-271
• /
• 1997

Ammonia removal capacities of five submerged filter media, 2~3mm sand, 30~50mm gravel, 20~40mm coral sand, polythylene net, and corrugated plastic plate in a seawater recirculating system were tested. A rotating biological contactor (RBC) was also tested for comparison. Oxygen consumption rates were measured along with the ammonia removal efficiencies. The ammonia concentrations in the system were maintained from 0.052 to 0.904 mg/l (mean 0.338$\pm$0.219 mg/l) and the water temperature was ranged from 19.2 to $21.4^{\circ}C\;(mean 20.2^{\circ}C\pm0.58^{\circ}C$). The 1/2-order kinetic model (Y:g/$m^3$/day) and the mean ammonia removal rates (g/$m^3$/day) of the filter media were : Sand : Y=135.5X0.5-25.1(r2=0.8110), 45.1 Coral sand : Y=125.1X0.5-33.0 (r2=0.7307), 31.8 Polyethylene net : Y=87.4X0.5-20.1 (r2=0.6780), 25.2 Corrugated plastic plate : Y=87.4X0.5-20.1(r2=0.5206), 19.2 Gravel : Y=4307X0.5-5.5 (r2=0.2596), 17.1 RBC : Y=127.6X0.5-33.4 (r2=0.7146), 32.8 where X is the concentration of ammonia. Oxygen consumption rates well corresponded to the ammonia removal capacities of each filter medium, thus the sands showing the highest value (442g/$m^3$/day) followed by coral sands (291.1g/$m^3$/day), polyethylene nets (236.9g/$m^3$/day), gravels (135.6g/$m^3$/day) and corrugated plastic plates (134.2g/$m^3$/day). Oxygen consumption rate of the RBC was unable to measure because of the characteristics of the structure.

• Korean Journal of Occupational Health Nursing
• /
• v.7 no.1
• /
• pp.5-19
• /
• 1998

This descriptive-correlational study was conducted to identify the major factors affecting health promoting behaviors. 344 workers who employed in four different manutacturing plants in Taegu and Kyungbuk area were selected by convenience sampling method. Data were collected from April let to April 18th, 1998 by ready structured questionaires. The purpose of this study was to offer the basic data for health promotion theory development and health promotion strategy planning. This study was based on Pender's Health Promotion Model and examined three variables health promoting behavior, self-concept and perceived importance of health. The Life Style and Health Habit Assessment scale(LHHA) developed by Pender(1982).The Self-concept scale developed by Choi(1972) and the Health Value scale developed by Wallston, Maides and Wallston(1980) were used for this study. Data was analyzed by percentage, mean. t-test. ANOVA, Pearson Correlation Coefficient, and Stepwise Multiple Regression. The major findings of this study are as follows ; 1. The average level of health promoting behavior practice was 63.2% and possible range was from 62 to 248 point. The mean score of respondent's positive self-concept was 75.8. 81.4% of respondents put a high priority on the importance of health. 2. There was a significant difference between the practice level in the category of general self care and less amount of working hours per day(P=0.000), less amount of working hours per week(P=0.000). There was a significant difference between the practice level in the category of nutrition and age(0.002), marital status(0.000), working hour per day(0.008), working hours per week(0.001), There was a significant difference between the practice level in the category of nutriton and sex(0.000), age(0.000), marital status(0.025), education level(0.000), working hours per day(0.002), working hours per week(0.006). There was a significant difference between the practice level in the category of sleep and rest and age(0.003), marital status(0.002), working hours per day(0.001), working hours per week(0.001). There was a significant difference between the practice level in the category of stress management and working hours per day(0.001), working hours per week(0.002). There was a significant difference between the practice level in the category of self-actualization and working hours per day(0.050). 3. General characteristics influencing the respodent's self-concept were level(P=0.009) and worksite(P=0.001). 4. The results of the hypothesis tests are as follows The first hypothesis, that "The respondent who have more positive self-concept will have higher scores in the practice of health promoting behavior." was supported(r=0.2973, P=0.0001). The second hypothesis that "The respondent who have higher perception level on importance of health will have higher scores in the practice health promoting behavior." was rejected(r=- 0665, P=0.2225). 5. The most important factor that affects health promoting behavior practice was working hours per week(6.0%). The combination of working hours per week, age, education level accounted for 10.0% of the variance in health promoting behavior. In conclusion, the results of this study on industrial workers supported Pender's health promotion model in partial and showed the relatedness between self concept and the practice of health promoting behavior. Further research is required to find factors influencing health promoting behaviors of industrial workers.

• Journal of Preventive Medicine and Public Health
• /
• v.32 no.2
• /
• pp.177-182
• /
• 1999

Objectives: To reexamine the association between air pollution and daily mortality in Seoul, Korea using a method of meta-analysis with the data filed for 1991 through 1995. Methods: A separate Poisson regression analysis on each district within the metropolitan area of Seoul was conducted to regress daily death counts on levels of each ambient air pollutant, such as total suspended particulates (TSP), sulfur dioxide $(SO_2)$, and ozone $(O_3)$, controlling for variability in the weather condition. We calculated a weighted mean as a meta-analysis summary of the estimates and its standard error. Results: We found that the p value from each pollutant model to test the homogeneity assumption was small (p<0.01) because of the large disparity among district-specific estimates. Therefore, all results reported here were estimated from the random effect model. Using the weighted mean that we calculated, the mortality at a $100{\mu}g/m^3$ increment in a 3-day moving average of TSP levels was 1.034 (95% Cl 1.009-1.059). The mortality was estimated to increase 6% (95% Cl 3-10%) and 3% (95% Cl 0-6%) with each 50 ppb increase for 9-day moving average of SO2 and 1-hr maximum O3, respectively. Conclusions: Like most of air pollution epidemiologic studies, this meta-analysis cannot avoid fleeing from measurement misclassification since no personal measurement was taken. However, we can expect that a measurement bias be reduced in a district-specific estimate since a monitoring station is hefter representative cf air quality of the matched district. The similar results to those from the previous studios indicated existence of health effect of air pollution at current levels in many industrialized countries, including Korea.

• Korean Journal of Ecology and Environment
• /
• v.47 no.3
• /
• pp.160-166
• /
• 2014

This study was aimed at looking into the distribution status and age structure of Abies holophylla population in Sudo-Am temple forest. It was found that a total of 302 individuals of Abies holophylla existed which were more than 2m in height within the study area. Furthermore the population size is one of the largest in the southern region of Korea. The CBH of Abies holophylla ranged from 1.5 cm to 500.8 cm. Age structure of Abies holophylla looks like a gourd-shaped bottle. This means that they have an unstable structure status and do not survive very long. This status results from a variety of factors including, vegetation succession, anthropogenic activities, and global warming. The environmental characteristics of Abies holophylla population was $931{\pm}64.5m$ in mean altitude, $19.2{\pm}8.7^{\circ}$ in mean slope in the northeastern and southeastern area of the slope direction, and $1,324,323{\pm}174,459wh\;m^{-2}$ in average of direct normal irradiation. Among the site environmental factors, the significant ones which influence the potential habitat for Abies holophylla distribution were chosen using the MaxEnt model. According to the results of this study, altitude and slope were found as the important factors. The average value of environmental conditions by ROC analysis were altitude 903.2 m, slope $20.04^{\circ}$, irradiation $1,352.248wh\;m^{-2}$, and the southeastern aspect.