• 융합정보논문지
• /
• 제9권8호
• /
• pp.1-8
• /
• 2019

대화시스템은 인간과 컴퓨터의 상호작용에 새로운 패러다임이 되고 있다. 자연어로써 상호작용함으로써 인간은 보다 자연스럽고 편리하게 각종 서비스를 누릴 수 있게 되었다. 대화시스템의 구조는 일반적으로 음성 인식, 자연어 이해, 문맥 파악 등의 여러 모듈의 파이프라인으로 이뤄지는데, 본 연구에서는 자연어 이해 모듈의 도메인 분류 문제를 풀기 위해 convolutional neural network, random forest 등의 기계학습 모델을 비교하였다. 사람이 직접 태깅한 총 7개 서비스 도메인 데이터에 대하여 각 문장의 도메인을 분류하는 실험을 수행하였고 random forest 모델이 F1 score 0.97 이상으로 가장 높은 성능을 달성한 것을 보였다. 향후 다른 기계학습 모델들을 추가 실험함으로써 도메인 분류 성능 개선을 지속할 계획이다.

• 한국공간구조학회논문집
• /
• 제21권2호
• /
• pp.99-110
• /
• 2021

This study presents the estimation of crack depth by analyzing temperatures extracted from thermal images and environmental parameters such as air temperature, air humidity, illumination. The statistics of all acquired features and the correlation coefficient among thermal images and environmental parameters are presented. The concrete crack depths were predicted by four different machine learning models: Multi-Layer Perceptron (MLP), Random Forest (RF), Gradient Boosting (GB), and AdaBoost (AB). The machine learning algorithms are validated by the coefficient of determination, accuracy, and Mean Absolute Percentage Error (MAPE). The AB model had a great performance among the four models due to the non-linearity of features and weak learner aggregation with weights on misclassified data. The maximum depth 11 of the base estimator in the AB model is efficient with high performance with 97.6% of accuracy and 0.07% of MAPE. Feature importances, permutation importance, and partial dependence are analyzed in the AB model. The results show that the marginal effect of air humidity, crack depth, and crack temperature in order is higher than that of the others.

• 국토계획
• /
• 제54권3호
• /
• pp.106-118
• /
• 2019

In order to develop pedestrian navigation service that provides optimal pedestrian routes based on pedestrian satisfaction levels, it is required to develop a prediction model that can estimate a pedestrian's satisfaction level given a certain condition. Thus, the aim of the present study is to develop a pedestrian satisfaction prediction model based on three machine learning algorithms: Logistic Regression, Random Forest, and Artificial Neural Network models. The 2009, 2012, 2013, 2014, and 2015 Pedestrian Satisfaction Survey Data in Seoul, Korea are used to train and test the machine learning models. As a result, the Random Forest model shows the best prediction performance among the three (Accuracy: 0.798, Recall: 0.906, Precision: 0.842, F1 Score: 0.873, AUC: 0.795). The performance of Artificial Neural Network is the second (Accuracy: 0.773, Recall: 0.917, Precision: 0.811, F1 Score: 0.868, AUC: 0.738) and Logistic Regression model's performance follows the second (Accuracy: 0.764, Recall: 1.000, Precision: 0.764, F1 Score: 0.868, AUC: 0.575). The precision score of the Random Forest model implies that approximately 84.2% of pedestrians may be satisfied if they walk the areas, suggested by the Random Forest model.

• Computers and Concrete
• /
• 제30권1호
• /
• pp.43-74
• /
• 2022

Machine learning technique is recently opening new opportunities to identify the complex shear transfer mechanisms of reinforced concrete (RC) beam members. This study employed 1224 shear test specimens to train decision tree-based machine learning (ML) programs, by which strong correlations between shear capacity of RC beams and key input parameters were affirmed. In addition, shear contributions of concrete and shear reinforcement (the so-called V_c and V_s) were identified by establishing three independent ML models trained under different strategies with various combinations of datasets. Detailed parametric studies were then conducted by utilizing the well-trained ML models. It appeared that the presence of shear reinforcement can make the predicted shear contribution from concrete in RC beams larger than the pure shear contribution of concrete due to the intervention effect between shear reinforcement and concrete. On the other hand, the size effect also brought a significant impact on the shear contribution of concrete (V_c), whereas, the addition of shear reinforcements can effectively mitigate the size effect. It was also found that concrete tends to be the primary source of shear resistance when shear span-depth ratio a/d<1.0 while shear reinforcements become the primary source of shear resistance when a/d>2.0.

• Advances in concrete construction
• /
• 제14권1호
• /
• pp.15-34
• /
• 2022

The use of waste rubber in concrete can reduce natural aggregate consumption and improve some technical properties of concrete. Although there are several equations for estimating the mechanical properties of concrete containing waste rubber, limited numbers of machine learning-based models have been proposed to predict the mechanical properties of rubbercrete. In this study, an extensive database of the mechanical properties of rubbercrete was gathered from a comprehensive survey of the literature. To model the mechanical properties of rubbercrete, M5P tree and linear gene expression programming (LGEP) methods as two machine learning techniques were employed to achieve reliable mathematical equations. Two procedures of input variable selection were considered in this study. The crucial component ratios of rubbercrete and concrete age were assumed as the input variables in the first procedure. In contrast, the volumes of the coarse and fine waste rubber and the compressive strength of concrete without waste rubber were considered the second procedure of the input variables. The results show that the models obtained by LGEP are more accurate than those achieved by the M5P model tree and existing traditional equations. Besides, the volumes of the coarse and fine waste rubber and the compressive strength of concrete without waste rubber are better predictors of the mechanical properties of rubbercrete compared to the first procedure of input variable selection.

• Journal of Forest and Environmental Science
• /
• 제39권4호
• /
• pp.195-202
• /
• 2023

In this study, we sought to compare and evaluate the accuracy and predictive performance of machine learning algorithms for estimating the growth of individual Larix kaempferi trees in Gangwon Province, Korea. We employed linear regression, random forest, XGBoost, and LightGBM algorithms to predict tree growth using monitoring data organized based on different thinning intensities. Furthermore, we compared and evaluated the goodness-of-fit of these models using metrics such as the coefficient of determination (R²), mean absolute error (MAE), and root mean square error (RMSE). The results revealed that XGBoost provided the highest goodness-of-fit, with an R² value of 0.62 across all thinning intensities, while also yielding the lowest values for MAE and RMSE, thereby indicating the best model fit. When predicting the growth volume of individual trees after 3 years using the XGBoost model, the agreement was exceptionally high, reaching approximately 97% for all stand sites in accordance with the different thinning intensities. Notably, in non-thinned plots, the predicted volumes were approximately 2.1 m³ lower than the actual volumes; however, the agreement remained highly accurate at approximately 99.5%. These findings will contribute to the development of growth prediction models for individual trees using machine learning algorithms.

• Design & Manufacturing
• /
• 제18권2호
• /
• pp.64-73
• /
• 2024

The purpose of this paper is to introduce a fusion method that combines the design of experiments (DOE) and machine learning to optimize the bias of plastic products. The study focuses on the plastic motor housing used in automobiles, which is manufactured through plastic injection molding. Achieving optimal molding for the motor housing involves the optimization of various molding conditions, including injection pressure, injection time, holding pressure, mold temperature, and cooling time. Failure to optimize these conditions can lead to increased product deformation. To minimize the deformation of the motor housing, the widely used Taguchi method, which is one of the design of experiment techniques, was employed to identify the injection molding conditions that affect deformation. Machine learning was then applied to various models based on the identified molding conditions. Among the models, the Random Forest model emerged as the most effective in predicting deformation amounts. The validity of the Random Forest model was also confirmed through verification. The verification results demonstrated the excellent prediction accuracy of the trained Random Forest model. By utilizing the validated model, molding conditions that minimize deformation were determined. Implementation of these optimal molding conditions led to a reduction of approximately 5.3% in deformation compared to the conditions before optimization. It is noteworthy that all injection molding outcomes presented in this paper were obtained through robust injection molding simulations, ensuring both research objectivity and speed.

• ETRI Journal
• /
• 제44권4호
• /
• pp.613-623
• /
• 2022

Autism spectrum disorder (ASD) is a developmental disorder associated with cognitive and neurobehavioral disorders. It affects the person's behavior and performance. Autism affects verbal and non-verbal communication in social interactions. Early screening and diagnosis of ASD are essential and helpful for early educational planning and treatment, the provision of family support, and for providing appropriate medical support for the child on time. Thus, developing automated methods for diagnosing ASD is becoming an essential need. Herein, we investigate using various machine learning methods to build predictive models for diagnosing ASD in children using facial images. To achieve this, we used an autistic children dataset containing 2936 facial images of children with autism and typical children. In application, we used classical machine learning methods, such as support vector machine and random forest. In addition to using deep-learning methods, we used a state-of-the-art method, that is, automated machine learning (AutoML). We compared the results obtained from the existing techniques. Consequently, we obtained that AutoML achieved the highest performance of approximately 96% accuracy via the Hyperpot and tree-based pipeline optimization tool optimization. Furthermore, AutoML methods enabled us to easily find the best parameter settings without any human efforts for feature engineering.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제9권8호
• /
• pp.3169-3181
• /
• 2015

Nowadays, online word-of-mouth has become a powerful influencer to marketing and sales in business. Opinion mining and sentiment analysis is frequently adopted at market research and business analytics field for analyzing word-of-mouth content. However, there still remain several challengeable areas for 1) sentiment analysis aiming for Korean word-of-mouth content in film market, 2) availability of machine learning models only using linguistic features, 3) effect of the size of the feature set. This study took a sample of 10,000 movie reviews which had posted extremely negative/positive rating in a movie portal site, and conducted sentiment analysis with four machine learning algorithms: naïve Bayesian, decision tree, neural network, and support vector machines. We found neural network and support vector machine produced better accuracy than naïve Bayesian and decision tree on every size of the feature set. Besides, the performance of them was boosting with increasing of the feature set size.

• Journal of information and communication convergence engineering
• /
• 제21권4호
• /
• pp.281-286
• /
• 2023

Diagnosing reading disorders involves complex procedures to evaluate complex cognitive processes. For an accurate diagnosis, a series of tests and evaluations by human experts are required. In this study, we propose a quantitative tool to diagnose reading disorders based on natural reading behaviors using minimal human input. The eye movements of the third- and fourth-grade students were recorded while they read a text at their own pace. Seven machine learning models were used to evaluate the gaze patterns of the words in the presented text and classify the students as normal or having a reading disorder. The accuracy of the machine learning-based diagnosis was measured using the diagnosis by human experts as the ground truth. The highest accuracy of 0.8 was achieved by the support vector machine and random forest classifiers. This result demonstrated that machine learning-based automated diagnosis could substitute for the traditional diagnosis of reading disorders and enable large-scale screening for students at an early age.