• Journal of Nutrition and Health
• /
• v.46 no.1
• /
• pp.50-60
• /
• 2013

The aims of this study were to estimate daily intake of macrominerals from beverages, liquid teas, and liquid coffees and to evaluate their potential health risks for Korean children and adolescents (1-to 19 years old). Assessment of dietary intake was conducted using the actual level of sodium, calcium, phosphorus, potassium, and magnesium in non-alcoholic beverages and (207 beverages, 19 liquid teas, and 24 liquid coffees) the food consumption amount drawn from "The Fourth Korea National Health and Nutrition Examination Survey (2007-2009)". To estimate the dietary intake of non-alcoholic beverages, 6,082 children and adolescents (Scenario I) were compared with 1,704 non-alcoholic beverage consumption subjects among them (Scenario II). Calculation of the estimated daily intake of macrominerals was based on point estimates and probabilistic estimates. The values of probabilistic macromineral intake, which is a Monte-Carlo approach considering probabilistic density functions of variables, were presented using the probabilistic model. The level of safety for macrominerals was evaluated by comparison with population nutrient intake goal (Goal, 2.0 g/day) for sodium, tolerable upper intake level (UL) for calcium (2,500 mg/day) and phosphorus (3,000-3,500 mg/day) set by the Korean Nutrition Society (Dietary Reference Intakes for Koreans, KDRI). For total children and adolescents (Scenario I), mean daily intake of sodium, calcium, phosphorus, potassium, and magnesium estimated by probabilistic estimates using Monte Carlo simulation was, respectively, 7.93, 10.92, 6.73, 23.41, and 1.11, and 95th percentile daily intake of those was, respectively, 28.02, 44.86, 27.43, 98.14, and 3.87 mg/day. For consumers-only (Scenario II), mean daily intake of sodium, calcium, phosphorus, potassium, and magnesium estimated by probabilistic estimates using Monte Carlo simulation was, respectively, 19.10, 25.77, 15.83, 56.56, and 2.86 mg/day, and 95th percentile daily intake of those was, respectively, 62.67, 101.95, 62.09, 227.92, and 8.67 mg/day. For Scenarios I II, sodium, calcium, and phosphorus did not have a mean an 95th percentile intake that met or exceeded the 5% of Goal and UL.

• Journal of Intelligence and Information Systems
• /
• v.24 no.1
• /
• pp.167-181
• /
• 2018

Over the past decade, deep learning has been in spotlight among various machine learning algorithms. In particular, CNN(Convolutional Neural Network), which is known as the effective solution for recognizing and classifying images or voices, has been popularly applied to classification and prediction problems. In this study, we investigate the way to apply CNN in business problem solving. Specifically, this study propose to apply CNN to stock market prediction, one of the most challenging tasks in the machine learning research. As mentioned, CNN has strength in interpreting images. Thus, the model proposed in this study adopts CNN as the binary classifier that predicts stock market direction (upward or downward) by using time series graphs as its inputs. That is, our proposal is to build a machine learning algorithm that mimics an experts called 'technical analysts' who examine the graph of past price movement, and predict future financial price movements. Our proposed model named 'CNN-FG(Convolutional Neural Network using Fluctuation Graph)' consists of five steps. In the first step, it divides the dataset into the intervals of 5 days. And then, it creates time series graphs for the divided dataset in step 2. The size of the image in which the graph is drawn is $40(pixels){\times}40(pixels)$, and the graph of each independent variable was drawn using different colors. In step 3, the model converts the images into the matrices. Each image is converted into the combination of three matrices in order to express the value of the color using R(red), G(green), and B(blue) scale. In the next step, it splits the dataset of the graph images into training and validation datasets. We used 80% of the total dataset as the training dataset, and the remaining 20% as the validation dataset. And then, CNN classifiers are trained using the images of training dataset in the final step. Regarding the parameters of CNN-FG, we adopted two convolution filters ($5{\times}5{\times}6$ and $5{\times}5{\times}9$) in the convolution layer. In the pooling layer, $2{\times}2$ max pooling filter was used. The numbers of the nodes in two hidden layers were set to, respectively, 900 and 32, and the number of the nodes in the output layer was set to 2(one is for the prediction of upward trend, and the other one is for downward trend). Activation functions for the convolution layer and the hidden layer were set to ReLU(Rectified Linear Unit), and one for the output layer set to Softmax function. To validate our model - CNN-FG, we applied it to the prediction of KOSPI200 for 2,026 days in eight years (from 2009 to 2016). To match the proportions of the two groups in the independent variable (i.e. tomorrow's stock market movement), we selected 1,950 samples by applying random sampling. Finally, we built the training dataset using 80% of the total dataset (1,560 samples), and the validation dataset using 20% (390 samples). The dependent variables of the experimental dataset included twelve technical indicators popularly been used in the previous studies. They include Stochastic %K, Stochastic %D, Momentum, ROC(rate of change), LW %R(Larry William's %R), A/D oscillator(accumulation/distribution oscillator), OSCP(price oscillator), CCI(commodity channel index), and so on. To confirm the superiority of CNN-FG, we compared its prediction accuracy with the ones of other classification models. Experimental results showed that CNN-FG outperforms LOGIT(logistic regression), ANN(artificial neural network), and SVM(support vector machine) with the statistical significance. These empirical results imply that converting time series business data into graphs and building CNN-based classification models using these graphs can be effective from the perspective of prediction accuracy. Thus, this paper sheds a light on how to apply deep learning techniques to the domain of business problem solving.

• The korean journal of orthodontics
• /
• v.25 no.4
• /
• pp.465-485
• /
• 1995

The purpose of this study is to locate the proper position of the lower occlusal plane according to individual skeletal pattern. Cephalometric films of 234 subjects of the control group, 358 of the pretreatment group and 358 of the treated group were analyzed to study proper relationships between vertical dimension ratio(VDR) and lower occlusomandibular plane angle(LOM). The control group was divided into two subgroups by the age. The first subgroup consisted of 113 subjects of the age 14 years and under and with the mean age of 10.82 years. The other subgroup consisted of 113 subjects of the age 18 years and above with the mean age of 23.76 years. The pretreatment group was divided into three subgroups by the age. The first subgroup consisted of 274 subjects of the age 14 years and under with the mean age of 11.36 years. The second subgroup consisted of 54 subjects of the age 14 through 18 years with the mean age of 15.4 years. The last subgroup consisted of 30 subjects of the age 18 years and above with the mean age of 21.35 years. The treated group was also divided into three subgroups by the age. The first subgroup consisted of 145 subjects of the age 14 years and under with the mean age of 12.91 years. The second subgroup consisted of 166 subjects of the age 14 through 18 years with the mean age of 15.64 years. The last subgroup consisted of 47 subjects of the age 18 years and above with the mean age of 21.61 years. Cephalometric films of the sample were traced. Measurements were made to a hundredth using a program specifically prepared for this study, and the results were entered into a 486DX PC. Means and Standard deviations of all the veriables were calculated for each group. Correlation coefficients between pertinent variables were calculated. Significance tests on those coefficients, one-way ANOVA and t-tests between variables or groups were performed. On the basis of the results studied above, certain subjects were selected from the control and the treated groups to locate the proper position of the occlusal plane, and designated as the optimal occluaion group. The subjects of this optimal occlusion group had 1-3 mm overbite, 1-3 mm of overjet and less than 1.75 mm of curve of Spee. A total subjects of 187 in this group consisted 104 treated subjects and 83 control group. Regression analysis was carried out between VDR and LOM, and regression equations were tabulated for this optimal occlusion group. The results were as follows : 1. Highly significant correlations were observed between various variables useful for identifying vertical component of skeletal frame, but any one particular variable did not accurately indicate the magnitude of anterior vertical overbite. 2. Of the variables useful identifying vertical component of skeletal frame, The VDR showed the highest correlation to the LOM. 3. Of the total sample, 80 percent had overbite within the normal range, irrespective of VDR. 4. The optimal occlusion group was divided into 9 subgroups by the age and the anteroposterior skeletal pattern, and correlation coefficient and determination coefficient between VDR and LOM of each group were calculated. Correlation coefficients and determination coefficients were found to be significantly high in all groups. 5. Regression equation was induced for each of the optimal occlusion group to find proper LOM according to the VDR. 6. It was found that the mean value of the cant of occlusal plane itself is not enough for a diagnosis and a treatment plan. Rather, It is very important to locate the proper occlusal plane for an Individual skeletal pattern.

• Radiation Oncology Journal
• /
• v.15 no.4
• /
• pp.349-356
• /
• 1997

Purpose : To evaluate the efficacy of hyperfractionated radiation therapy in carcinoma of the cervix, especially on huge exophytic and endophytic stage Ib, IIa and IIb Materials and Materials : Fourty one patients with carcinoma of the cervix treated with hyperfractionated radiation therapy at the Department of Therapeutic Radiology, Dongsan Hospital, Keimyung University. School of Medicine from Jul, 1991 to Apr, 1994. According to FIGO s1aging system, therewere stage Ib (3 patients) IIa (6 patients) with exophytic ($\geq$5cm in dinmeter) and huge endophytic mass. and IIb (32 patients) with median age of 55 yeavs old. Radiation therapy consisted of hyperfractionated external irradition to the whole pelvis (120cGy/fraction, 2 fraction/day (minimum interval of 6 hours), 3600-5520cGy) and boost parametrial doses (for a total of 4480-6480cGy) with midline shield $(4\times10cm)$, and combined with intracavitary irradiation (up to 7480-8520cGy in Ib, IIa and 8480-9980cGy in IIb to point A). The maximum and mean follow up durations were 70 and 47 months respectively . Results : Five year local control rate was $78\%$ and the actuarial overall five year survival rate was $66.1\%$ for all patients, $44.4\%$ for stage Ib, IIa and $71.4\%$ for stage IIb. In bulky IIb (above 5cm in tumor size, 11 patients) five year local control rate and five rear survival rate was $88.9\%,\;73\%$ respectively Pelvic lymph node status (negative : $74\%,\;positive:25\%$, p=0.0015) was significant Prognostic factor affecting to five rear survival rate. There was marginally significant survival difference by total dose to A point ($>84Gy\;:\;70\%,\;>84Gy\;:\;42.8\%$, p=0.1). We consider that the difference of total dose to A point by stage (mean Ib,IIa : 79Gy. IIb 89Gy P=0.001) is one of the causes in worse local control and survival of Ib,IIa than IIb The overall recurrence rate was $39\%$ (16/41). The rates of local failure alone. distant failure alone. and combined local and distant failure were $9.7\%,\;19.5\%,\;and\;9.7\%$, respectively. Two Patients developed leukopenia ($\geq$ grade 3) and Three patients develoued grade 3 gastrointestinal complication. Above grade 3 complication was not noted. There was no treatment related death noted. Conclusion : We thought that it may be necessary to increase A point dose to more than 85Gy in hyperfractionated radiotherapy of huge exophytic and endophvtic stage Ib,IIa. We considered that hyperfractionated radiation therapy may be tolerable in huge exophytic and endophytic stage IIb cervical carcinoma with acceptable morbidity and possible survival gain but this was results in small patient group and will be confirmed by long term follow up in many patients.

• Tuberculosis and Respiratory Diseases
• /
• v.43 no.5
• /
• pp.709-719
• /
• 1996

Background: Surgical resection is the only way to cure non-small cell lung cancer(NSCLC) and the prognosis of NSCLC in patients who undergo a complete resection is largely influenced by the pathologic stage. After surgical resection, recurrences in distant sites is more common than local recurrences. An effective postoperative adjuvant therapy which can prevent recurrences is necessary to improve long tenn survival Although chemotherapy and radiotherapy are still the mainstay in adjuvant therapy, the benefits of such therapies are still controversial. We initiated this retrospective study to evaluate the effects of adjuvant therapies and analyze the prognostic factors for survival after curative resection. Method: From 1990 to 1995, curative resection was perfomled in 282 NSCLC patients with stage I, II, IIIa, Survival analysis of 282 patients was perfonned by Kaplan-Meier method. The prognostic factors, affecting survival of patients were analyzed by Cox regression model. Results: Squamous cell carcinoma was present in 166 patients(59%) ; adenocarcinoma in 86 pmients(30%) ; adenosquamous carcinoma in II parients(3.9%); and large cell undifferentiated carcinoma in 19 patients(7.1%). By TNM staging system, 93 patients were in stage I; 58 patients in stage II ; and 131 patients in stage rna. There were 139 postoperative recurrences which include 28 local and 111 distant failures(20.1% vs 79.9%). The five year survival rate was 50.1% in stage I ; 31.3% in stage II ; and 24.1% in stage IIIa(p <0.0001). The median survival duration was 55 months in stage I ; 27 months in stage II ; and 16 months in stage rna. Among 131 patients with stage rna, the median survival duration was 19 months for 81 patients who received postoperative adjuvant chemotherapy only or cherne-radiotherapy and 14 months for the other 50 patients who received surgery only or surgery with adjuvant radiotherapy(p=0.2982). Among 131 patients with stage IIIa, the median disease free survival duration was 16 months for 21 patients who received postop. adjuvant chemotherapy only and 4 months for 11 patients who received surgery only(p=0.0494). In 131 patients with stage IIIa, 92 cases were in N2 stage. The five year survival rate of the 92 patients with N2 was 25% and their median survival duration was 15 months. The median survival duration in patients with N2 stage was 18 months for those 62 patients who received adjuvant chemotherapy and 14 months for the other 30 patients who did not(p=0.3988). The median survival duration was 16 months for those 66 patients who received irradiation and 14 months for the other 26 patients who did not(p=0.6588). We performed multivariate analysis to identify the factors affecting prognosis after complete surgical resection, using the Cox multiple regression model. Only age(p=0.0093) and the pathologic stage(p<0.0001) were significam prognostic indicators. Conclusion: The age and pathologic stage of the NSCLC parients are the significant prognostic factors in our study. Disease free survival duration was prolonged with statistical significance in patients who received postoperative adjuvant chemotherapy but overall survival duration was not affected according to adjuvant therapy after surgical resection.

• Journal of the Korean Society of Food Science and Nutrition
• /
• v.43 no.8
• /
• pp.1304-1316
• /
• 2014

The aims of this study were to estimate daily intakes of artificial sweeteners from beverages and liquid teas as well as evaluate their potential health risks in Korean children and adolescents (1 to 19 years old). Dietary intake assessment was conducted using actual levels of aspartame, acesulfame-K, and sucralose in non-alcoholic beverages (651 beverages and 87 liquid teas), and food consumption amounts were drawn from "The Fourth Korea National Health and Nutrition Examination Survey (2007~2009)". To estimate dietary intake of non-alcoholic beverages, a total of 6,082 children and adolescents (Scenario I) were compared to 1,704 non-alcoholic beverage consumption subjects (Scenario II). The estimated daily intake of artificial sweeteners was calculated based on point estimates and probabilistic estimates. The values of probabilistic artificial sweeteners intakes were presented by a Monte Carlo approach considering probabilistic density functions of variables. The level of safety for artificial sweeteners was evaluated by comparisons with acceptable daily intakes (ADI) of aspartame (0~40 mg/kg bw/day), acesulfame-K (0~15 mg/kg bw/day), and sucralose (0~15 mg/kg bw/day) set by the World Health Organization. For total children and adolescents (Scenario I), mean daily intakes of aspartame, acesulfame-K, and sucralose estimated by probabilistic estimates using Monte Carlo simulation were 0.09, 0.01, and 0.04 mg/kg bw/day, respectively, and 95th percentile daily intakes were 0.30, 0.02, and 0.13 mg/kg bw/day, respectively. For consumers-only (Scenario II), mean daily intakes of aspartame, acesulfame-K, and sucralose estimated by probabilistic estimates using Monte Carlo simulation were 0.52, 0.03, and 0.22 mg/kg bw/day, respectively, and 95th percentile daily intakes were 1.80, 0.12, and 0.75 mg/kg bw/day, respectively. For scenarios I and II, neither aspartame, acesulfame-K, nor sucralose had a mean and 95th percentile intake that exceeded 5.06% of ADI.

• 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.