• Korean Journal of Remote Sensing
• /
• v.37 no.4
• /
• pp.719-731
• /
• 2021

This study proposes a two-stage hybrid classification model for crop classification using multi-temporal remote sensing images; the model combines feature embedding by using an autoencoder (AE) with a convolutional neural network (CNN) classifier to fully utilize features including informative temporal and spatial signatures. Long short-term memory (LSTM)-based AE (LAE) is fine-tuned using class label information to extract latent features that contain less noise and useful temporal signatures. The CNN classifier is then applied to effectively account for the spatial characteristics of the extracted latent features. A crop classification experiment with multi-temporal unmanned aerial vehicle images is conducted to illustrate the potential application of the proposed hybrid model. The classification performance of the proposed model is compared with various combinations of conventional deep learning models (CNN, LSTM, and convolutional LSTM) and different inputs (original multi-temporal images and features from stacked AE). From the crop classification experiment, the best classification accuracy was achieved by the proposed model that utilized the latent features by fine-tuned LAE as input for the CNN classifier. The latent features that contain useful temporal signatures and are less noisy could increase the class separability between crops with similar spectral signatures, thereby leading to superior classification accuracy. The experimental results demonstrate the importance of effective feature extraction and the potential of the proposed classification model for crop classification using multi-temporal remote sensing images.

• International Journal of Advanced Culture Technology
• /
• v.9 no.4
• /
• pp.226-236
• /
• 2021

The purpose of this study is to analyze the type of latent profile for general high school students' subject selection criteria and to identify the characteristics of the latent class. The survey data of 1072 general high school students (male; 648, female; 424) in G city, Jeollabuk-do and the scale composed of 8 sub-factors: 'SAT orientation', 'academic achievement', 'ability orientation', 'pursuit of interest', 'teacher orientation', 'career development', 'others' recommendation', and 'subject availability' were used for latent profile analysis and cross-analysis between potential layers. As a result of the analysis, high school students' perceptions of subject selection were classified into four latent profiles. The four groups were named 'High Perception Type', 'Low Perception Type', 'Self-Directed Type', and 'Stability-Oriented Type' according to their types. It was found that there was a difference between the latent classes in the importance and performance level of the subject selection criteria. These results can help identify the subject selection tendencies of latent groups in the operation of the 2015 revised curriculum and the 2025 high school credit system that emphasizes the student-centered course selection curriculum and they can also provide customized course selection guidance considering individual differences.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.18 no.2
• /
• pp.44-57
• /
• 2019

It is crucial to understand how people make decisions on mode choice and to accurately predict their behaviors in transportation planning. One of avenues for advancing modeling is, in particular, taking into account for taste heterogeneity in modeling that can incorporate different decision-making processes across group. In this study, we hypothesize that how people make decisions on mode choice would differ by destination in that land use characteristics are heterogeneous by zone even if zones are all in the same area. To this end, we apply Latent Class Modeling (LCM) to commute trips in Seoul by using 2010 household travel diary survey, investigate types of latent classes with the aid of characteristics of destination, and analyze how those classes differently response to factors. The LCM identifies two classes: in the first one, modal split of auto and public transit (bus and metro) is almost half-and-half and the trip destinations are characterized by relatively more residence facilities and less business/commercial facilities; in the second one, public transit has a notably high share and trip destinations are characterized by relatively more business/commercial facilities. In addition, it turns out that demographic and socio-economic variables affect mode choice differently by class.

• The Korean Journal of Applied Statistics
• /
• v.31 no.6
• /
• pp.721-732
• /
• 2018

Latent class analysis (LCA) is an important tool to explore unseen latent groups in multivariate categorical data. In practice, it is important to select a suitable set of variables because the inclusion of too many variables in the model makes the model complicated and reduces the accuracy of the parameter estimates. Dean and Raftery (Annals of the Institute of Statistical Mathematics, 62, 11-35, 2010) proposed a headlong search algorithm based on Bayesian information criteria values to choose meaningful variables for LCA. In this paper, we propose a new variable selection procedure for LCA by utilizing posterior probabilities obtained from each fitted model. We propose a new statistic to measure the adequacy of LCA and develop a variable selection procedure. The effectiveness of the proposed method is also presented through some numerical studies.

• Human Ecology Research
• /
• v.60 no.1
• /
• pp.53-68
• /
• 2022

Using a person-centered approach, the current study investigated latent profiles for the emotional and behavioral problems of students in sixth-grade in elementary school and second grade of middle school. The aim was to explore latent transition patterns and verify the factors affecting the transitions. The participants were 1,937 adolescents who responded to the 3^rd year (6^th grade of elementary school; Time 1), 4^th year (1^st grade of middle school), and 5^th year (2^nd grade of middle school; Time 2) of the Korean Children Youth Panel Study. Latent profile and latent transition analyses were performed. The results were as follows: first, the latent profile of emotional and behavioral problems changed from Time 1 to Time 2. The latent groups at Time 1 were classified into low, moderate, high, and externalizing-dominant, whereas at Time 2, five groups were identified: low, moderate, high, externalizing-dominant, and withdrawal-dominant. Second, transition analyses revealed that although 22.3-57.0% of latent groups remained unchanged, there were significant changes over time between groups, as a new group ('withdrawal-dominant') emerged in Time 2. Third, different factors influenced the latent profile transition of emotional and behavioral problems depending on the transition pattern. Higher levels of self-esteem, better relationships with peers and teachers, and lower levels of parental inconsistency meant emotional and behavioral problems had not worsened at Time 2. The results suggest that early interventions are needed during the transition from childhood to early adolescence.

• Environmental and Resource Economics Review
• /
• v.21 no.4
• /
• pp.921-945
• /
• 2012

Conditional Logit (CL) model is widely used since its model estimation and interpretation of results of the model is relatively easy, on the other hand, it has the limit of preference heterogeneity of respondents being not fully considered. In this study we used the two models, Mixed Logit (ML) Model and Latent Class Model (LCM) to explain preference heterogeneity of respondents for protection for Boryeong Dam wetland. As a result of the examination for heterogeneity in Boryeong city and six metropolitan areas, we found there was significant difference between two regions. While there was explicit preference heterogeneity within respondents in Boryeong city, we found little heterogeneity within respondents in six metropolitan areas. Thus in the case of six metropolitan areas, CL model can be used for parameter estimation while in the case of Boryeong city, WTP estimates are based on parameter estimates from ML model to reflect the heterogeneity within respondents. Additionally, ML model with interaction and 2-class LCM for respondents in Boryeong city were used to explain the sources of the heterogeneity. The ML model with interaction has advantage of explaining individual unobserved heterogeneity. However The comarison between these two models reflects the fact that LCM provided added information that was not conveyed in the ML model with interaction. Thus, Preference heterogeneity within respondents in this study may be better explained by class level through LCM rather than indiviual level through ML model.

• Proceedings of the KWS Conference
• /
• v.43
• /
• pp.283-285
• /
• 2004

In this paper an important class of problems in welding which come under the category of phase change is considered, Solidification and melting are important process in welding field. Phase change problems are accompanied by either absorption or release of thermal energy i,e, heat transfer process. This is complicated by the release or absorption of the latent heat of fusion at the solid-liquid interface. In this study the liberation of latent heat is taken in to account using fixed grid method. The numerical simulation and the finite element codes for the heat transfer analysis including the latent heat term has been developed based on this fixed grid method.

• The Journal of the Convergence on Culture Technology
• /
• v.9 no.6
• /
• pp.581-587
• /
• 2023

The purpose of this study was to analyze the change trajectory and latent class types of self-esteem in first to fourth grade elementary school students. For the purpose, the Korean Children's Panel data were analyzed using potential growth model and the growth mixture model. As the results, the linear change model was selected as the most appropriate model. The change trajectory was found to increase slightly as the grade increased. In addition, four latent class groups were derived through: 'high level-maintenance,' 'low level-increase,' 'high level-decrease,' and 'low level-maintenance.' Most children were in the 'high level-maintenance' group, followed by 'high level-decrease,' 'low level-increase,' and 'low level-maintenance' groups. Therefore, based on the results of the study, we suggest that educational institutions and local communities pay attention to trends in elementary school students' self-esteem and provide appropriate support for students in each class.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.19 no.4
• /
• pp.383-394
• /
• 2018

This study analyzed the latent profile for identifying the difference in the dependence on smartphone use among middle school students in the 1st grade using the Korean Children and Youth Panel Survey (KCYPS). From the result of this study, first the latent class was separated according to the type of dependence on smartphone use. Class 1 included the students (from fifth grade in elementary school) whose level of reliance on smartphone use was low. Class 2 was selected as the group whose level of reliance on smartphone was high. Secondly, in comparing class 2 to class 1, it was found that the students who have a high probability of being in class 1 were those whose fathers are high achievers, have high early self-esteem and less age attachment. Thirdly, the students in class 1 had a higher sense of neglect than those in class 2. Furthermore, the self-determination of the students in class 2 mediated the effect of neglect on the adolescents' dependence on smartphone use both directly and indirectly.

• Environmental and Resource Economics Review
• /
• v.19 no.4
• /
• pp.853-874
• /
• 2010

This study examines the acceptability of nuclear facilities using the latent class analysis. Even though, nuclear power is useful in terms of economic and energy security aspects, it is very difficult to expand the existing nuclear power plants or build a new one. Many studies analysed the cause of unacceptability of nuclear facilities but it has not been focused how large portion of people are divided pro and con. It is very important to know the distribution of people by the attitude toward nuclear facilities in order to meet the long term National Energy Plan. Through the latent class analysis with 1,025 respondents, people are classified into three groups(favor-class, support-class, opposition-class). The favor-class is the largest group which has moderate good attitudes toward the nuclear facilities in terms of economy, cleanness. and necessity but concerns a little about safety. The second largest group is the support-class which comprises 1/4 portion of people. The people in the class show the aggressive support for the nuclear facilities. 15% of the respondents belong to the opposition-class which show the negative attitudes to expansion of neclear facilities. In order to increase the acceptability of nuclear faculties, the most urgent work for the government to do is to less people's concern about nuclear safety.