• Composites Research
• /
• v.35 no.5
• /
• pp.317-321
• /
• 2022

In this study, the PIC design method with machine learning that automatically assigning different stacking sequences according to loading types was applied bumper beam. The input value and labels of the training data for applying machine learning were defined as coordinates and loading types of reference elements that are part of the total elements, respectively. In order to compare the 2D and 3D implementation method, which are methods of representing coordinate value, training data were generated, and machine learning models were trained with each method. The 2D implementation method is divided FE model into each face and generating learning data and training machine learning models accordingly. The 3D implementation method is training one machine learning model by generating training data from the entire finite element model. The hyperparameter were tuned to optimal values through the Bayesian algorithm, and the k-NN classification method showed the highest prediction rate and AUC-ROC among the tuned models. The 3D implementation method revealed higher performance than the 2D implementation method. The loading type data predicted through the machine learning model were mapped to the finite element model and comparatively verified through FE analysis. It was found that 3D implementation PIC bumper beam was superior to 2D implementation and uni-stacking sequence composite bumper.

• Journal of the Korean GEO-environmental Society
• /
• v.23 no.4
• /
• pp.5-10
• /
• 2022

Recently, ground subsidence has been continuously occurring in downtown areas, threatening the safety of citizens. Various underground facilities such as water and sewage pipelines and communication pipelines are buried under the road. It is reported that the cause of ground subsidence is the deterioration of various facilities and the reckless development of the underground. In particular, it is known that the biggest cause of ground subsidence is the aging of sewage pipelines. As an existing study related to this, several representative factors of sewage pipelines were selected and a study to predict the risk of ground subsidence through statistical analysis has been conducted. In this study, a data SET was constructed using the characteristics of OO city's sewage pipe characteristics and ground subsidence data, The data set constructed from the characteristics of the sewage pipe of OO city and the location of the ground subsidence was used. The goal of this study was to present a classification model for the occurrence of ground subsidence according to the characteristics of sewage pipes through machine learning. In addition, the importance of each sewage pipe characteristic affecting the ground subsidence was calculated.

• The Journal of Bigdata
• /
• v.7 no.1
• /
• pp.213-224
• /
• 2022

In logistics and distribution, Market Basket Analysis (MBA) is used as an important means to analyze the correlation between major sales products and to increase internal operational efficiency. In particular, the results of market basket analysis are used as important reference data for decision-making processes such as product purchase prediction, product recommendation, and product display structure in stores. With the recent development of e-commerce, the number of items handled by a single distribution and logistics company has rapidly increased, And the existing analytical methods such as Apriori and FP-Growth have slowed down due to the exponential increase in the amount of calculation and applied to actual business. There is a limit to examining important association rules to overcome this limitation, In this study, at the Main-Category level, which is the highest classification system of products, the utility item set mining technique that can consider the sales volume of products together was used to first select a group of products mainly sold together. Then, at the sub-category level, the types of products sold together were identified using FP-Growth. By using this sequential layer filtering technique, it may be possible to reduce the unnecessary calculations and to find practically usable rules for enhancing the effectiveness and profitability.

• Journal of Gastric Cancer
• /
• v.10 no.4
• /
• pp.162-167
• /
• 2010

Purpose: The incidence of lymph node metastasis has been reported to range from 2.6 to 4.8% in early stage gastric cancer with mucosal invasion (T1a cancer). Lymph node metastasis in early stage gastric cancer is known as an important predictive factor. We analyzed the prediction factors of lymph node metastasis in T1a cancer. Materials and Methods: A total of 9,912 patients underwent radical gastrectomy due to gastric cancer from October 1994 to July 2006 in the Department Of Surgery at Samsung Medical Center. We did a retrospective analysis of 2,524 patients of these patients, ones for whom the cancer was confined within the mucosa. Results: Among the 2,524 patients, 57 (2.2%) were diagnosed with lymph node metastasis, and of these, cancer staging was as follows: 41 were N1, 8 were N2, and 8 were N3a. Univariate analysis of clinicopathological factors showed that the following factors were significant predictors of metastasis: tumor size larger than 4 cm, the presence of middle and lower stomach cancer, poorly differentiated adenocarcinoma and signet-ring cell carcinoma, diffuse type cancer (by the Lauren classification), and lymphatic invasion. Multivariate analysis showed that lymphatic invasion and tumor larger than 4 cm were significant factors with P<0.001 and P=0.024, respectively. Conclusions: The frequency of lymph node metastasis is extremely low in early gastric cancer with mucosal invasion. However, when lymphatic invasion is present or the tumor is larger than 4 cm, there is a greater likelihood of lymph node metastasis. In such cases, surgical treatments should be done to prevent disease recurrence.

• Journal of Internet Computing and Services
• /
• v.24 no.5
• /
• pp.17-27
• /
• 2023

Today, as AI (Artificial Intelligence) technology develops and its practicality increases, it is widely used in various application fields in real life. At this time, the AI model is basically learned based on various statistical properties of the learning data and then distributed to the system, but unexpected changes in the data in a rapidly changing data situation cause a decrease in the model's performance. In particular, as it becomes important to find drift signals of deployed models in order to respond to new and unknown attacks that are constantly created in the security field, the need for lifecycle management of the entire model is gradually emerging. In general, it can be detected through performance changes in the model's accuracy and error rate (loss), but there are limitations in the usage environment in that an actual label for the model prediction result is required, and the detection of the point where the actual drift occurs is uncertain. there is. This is because the model's error rate is greatly influenced by various external environmental factors, model selection and parameter settings, and new input data, so it is necessary to precisely determine when actual drift in the data occurs based only on the corresponding value. There are limits to this. Therefore, this paper proposes a method to detect when actual drift occurs through an Anomaly analysis technique based on XAI (eXplainable Artificial Intelligence). As a result of testing a classification model that detects DGA (Domain Generation Algorithm), anomaly scores were extracted through the SHAP(Shapley Additive exPlanations) Value of the data after distribution, and as a result, it was confirmed that efficient drift point detection was possible.

• Journal of The Korean Association For Science Education
• /
• v.19 no.4
• /
• pp.528-541
• /
• 1999

In order to obtain data for developing an ideal science curriculum. four kinds of General Science textbooks based on the 6th curriculum were analyzed. Particularly inquiry activities were analyzed by Scientific Inquiry Evaluation Inventory(SIEI). The results are as follows: 1) The average number of inquiry activities in four kinds of textbooks is 115.5. And the number in each textbook is very diverse: textbook A contains 94 inquiry activities, textbook B 147. textbook C 100 and textbook D 121. 2) As for the number of inquiry activity scopes in four kinds of textbook. observation comes to 22, experiment 117, interpreting data 196, investigation 64, discussion 51, classification 4 and prediction 8. And then the conceptional inquiry activity is about 2.3 times as many as the inquiry experiment. 3) According to the analysis of each inquiry task by SIEI. textbook A has 268, textbook B 328, textbook C 207 and textbook D 304. 4) In the analysis of the structure of inquiry activity, the evaluation of the competition and cooperation scale shows more emphasis on common tasks. no pooled results(87.1 %). The discussion scale mostly consists of activities without discussion required among students(83.5%). The evaluation of openness scale shows more emphasis on activities with problems, procedures and answers presented(58.3%). In the evaluation of inquiry scope scale, the inquiry scope scale mostly has the activities to demonstrate or verify the contents of the text(66.9%). 5) As for the analysis of inquiry activities as a whole. The inquiry pyramid in four kinds of General Science textbooks shows the type I that emphasizes the inquiry activities in low level such as gathering and organizing data. The inquiry index in four kinds of textbooks is average 47.8, shows very high level (above 35).

• Journal of The Korean Association For Science Education
• /
• v.32 no.2
• /
• pp.236-264
• /
• 2012

The purpose of this study is to provide teachers with sub-components of inquiry skills and help them to give direct instructions on the skills to their students. Inquiry skills and strategies are considered by-products of science and inquiry instruction by most of the science teachers. On the other hand, much research shows that many students are not familiar with the way that they can use inquiry skills therefore direct instruction on the inquiry skills is needed. The lack of guidance on the sub-components for the inquiry skills, however, results in science teachers' ignorance of the inquiry skills. As shown in the previous studies which suggest that without teachers' guidance, students cannot acquire the intended skills, and it is necessary to inform science teachers of the necessity for direct instruction on the inquiry skills and strategy as well as give them the sub-components of the inquiry skills. On the basis of the results from the previous research on the inquiry skills, this study presents the sub-components of basic inquiry skills (observation, classification, measure, prediction, and reasoning) and integrated inquiry skills (problem recognition, hypothesis formulation, control of variables, data transformation, data interpretation, drawing conclusion, and generalization).

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.26 no.1B
• /
• pp.61-67
• /
• 2006

Various methods have been applied for the research to model the relationship between rainfall-runoff, which shows a strong nonlinearity. In particular, most researches to model the relationship between rainfall-runoff using artificial neural networks have used back propagation algorithm (BPA), Levenberg Marquardt (LV) and radial basis function (RBF). and They have been proved to be superior in representing the relationship between input and output showing strong nonlinearity and to be highly adaptable to rapid or significant changes in data. The theory of artificial neural networks is utilized not only for prediction but also for classifying the patterns of data and analyzing the characteristics of the patterns. Thus, the present study applied self?organizing map (SOM) based on Kohonen's network theory in order to classify the patterns of rainfall-runoff process and analyze the patterns. The results from the method proposed in the present study revealed that the method could classify the patterns of rainfall in consideration of irregular changes of temporal and spatial distribution of rainfall. In addition, according to the results from the analysis the patterns between rainfall-runoff, seven patterns of rainfall-runoff relationship with strong nonlinearity were identified by SOM.

• Journal of the Korea Convergence Society
• /
• v.12 no.9
• /
• pp.11-19
• /
• 2021

A software system is required to change during its life cycle due to various requirements such as adding functionalities, fixing bugs, and adjusting to new computing environments. Such program code modification should be considered as carefully as a new system development becase unexpected software errors could be introduced. In addition, when reusing open source programs, we can expect higher quality software if code changes of the open source program are predicted in advance. This paper proposes a Convolutional Neural Network (CNN)-based deep learning model to predict source code changes. In this paper, the prediction of code changes is considered as a kind of a binary classification problem in deep learning and labeled datasets are used for supervised learning. Java projects and code change logs are collected from GitHub for training and testing datasets. Software metrics are computed from the collected Java source code and they are used as input data for the proposed model to detect code changes. The performance of the proposed model has been measured by using evaluation metrics such as precision, recall, F1-score, and accuracy. The experimental results show the proposed CNN model has achieved 95% in terms of F1-Score and outperformed the multilayer percept-based DNN model whose F1-Score is 92%.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.17 no.6
• /
• pp.121-132
• /
• 2018

Weather is an important factor affecting roadway transportation in many aspects such as traffic flow, driver 's driving patterns, and crashes. This study focuses on the relationship between weather and road surface condition and develops a model to estimate the road surface condition using machine learning. A road surface sensor was attached to the probe vehicle to collect road surface condition classified into three categories as 'dry', 'moist' and 'wet'. Road geometry information (curvature, gradient), traffic information (link speed), weather information (rainfall, humidity, temperature, wind speed) are utilized as variables to estimate the road surface condition. A variety of machine learning algorithms examined for predicting the road surface condition, and a two - stage classification model based on 'Random forest' which has the highest accuracy was constructed. 14 days of data were used to train the model and 2 days of data were used to test the accuracy of the model. As a result, a road surface state prediction model with 81.74% accuracy was constructed. The result of this study shows the possibility of estimating the road surface condition using the existing weather and traffic information without installing new equipment or sensors.