• Asian-Australasian Journal of Animal Sciences
• /
• v.33 no.10
• /
• pp.1558-1565
• /
• 2020

Objective: The objective of this study was to characterize the number of loci affecting growth traits and the distribution of single nucleotide polymorphism (SNP) effects on growth traits, and to understand the genetic architecture for growth traits in Hanwoo (Korean cattle) using genome-wide association study (GWAS), genomic partitioning, and hierarchical Bayesian mixture models. Methods: GWAS: A single-marker regression-based mixed model was used to test the association between SNPs and causal variants. A genotype relationship matrix was fitted as a random effect in this linear mixed model to correct the genetic structure of a sire family. Genomic restricted maximum likelihood and BayesR: A priori information included setting the fixed additive genetic variance to a pre-specified value; the first mixture component was set to zero, the second to 0.0001×σ²_g, the third 0.001×σ²_g, and the fourth to 0.01×σ²_g. BayesR fixed a priori information was not more than 1% of the genetic variance for each of the SNPs affecting the mixed distribution. Results: The GWAS revealed common genomic regions of 2 Mb on bovine chromosome 14 (BTA14) and 3 had a moderate effect that may contain causal variants for body weight at 6, 12, 18, and 24 months. This genomic region explained approximately 10% of the variance against total additive genetic variance and body weight heritability at 12, 18, and 24 months. BayesR identified the exact genomic region containing causal SNPs on BTA14, 3, and 22. However, the genetic variance explained by each chromosome or SNP was estimated to be very small compared to the total additive genetic variance. Causal SNPs for growth trait on BTA14 explained only 0.04% to 0.5% of the genetic variance Conclusion: Segregating mutations have a moderate effect on BTA14, 3, and 19; many other loci with small effects on growth traits at different ages were also identified.

• Animal Bioscience
• /
• v.35 no.5
• /
• pp.648-658
• /
• 2022

Objective: The identification of nonlinear mixed models that describe the growth trajectory of New Zealand rabbits was performed based on weight records and carcass measures obtained using ultrasonography. Methods: Phenotypic records of body weight (BW) and loin eye area (LEA) were collected from 66 animals raised in a didactic-productive module of cuniculture located in the southern Piaui state, Brazil. The following nonlinear models were tested considering fixed parameters: Brody, Gompertz, Logistic, Richards, Meloun 1, modified Michaelis-Menten, Santana, and von Bertalanffy. The coefficient of determination (R²), mean squared error, percentage of convergence of each model (%C), mean absolute deviation of residuals, Akaike information criterion (AIC), and Bayesian information criterion (BIC) were used to determine the best model. The model that best described the growth trajectory for each trait was also used under the context of mixed models, considering two parameters that admit biological interpretation (A and k) with random effects. Results: The von Bertalanffy model was the best fitting model for BW according to the highest value of R² (0.98) and lowest values of AIC (6,675.30) and BIC (6,691.90). For LEA, the Logistic model was the most appropriate due to the results of R² (0.52), AIC (783.90), and BIC (798.40) obtained using this model. The absolute growth rates estimated using the von Bertalanffy and Logistic models for BW and LEA were 21.51g/d and 3.16 cm², respectively. The relative growth rates at the inflection point were 0.028 for BW (von Bertalanffy) and 0.014 for LEA (Logistic). Conclusion: The von Bertalanffy and Logistic models with random effect at the asymptotic weight are recommended for analysis of ponderal and carcass growth trajectories in New Zealand rabbits. The inclusion of random effects in the asymptotic weight and maturity rate improves the quality of fit in comparison to fixed models.

• Journal of the Society of Disaster Information
• /
• v.19 no.1
• /
• pp.161-172
• /
• 2023

Safety education delivered to construction workers in a lecture manner has limitations in concentration and immersion, so delivery power and interest are low. In order to improve unstable behavior through education and prevent safety accidents, it is necessary to change the paradigm to hands-on education. Purpose: Experiential safety education aims to contribute to preventing accidents for construction workers by quickly recognizing risks, improving emergency response skills, and verifying the effectiveness of pre- and post-learning. Method: Based on a survey of workers who experienced the same work environment as the actual construction site, an opinion survey on the pre- and post-safety experience education and a variable measurement tool were planned, and a research hypothesis was established. Results: The Bayesian theory and MC simulation analysis were used to analyze the structural equation model, and the change in construction worker behavior was confirmed through the intended safety (A), non-experiential education in the sub-area of anxiety (B), average, standard deviation, and minimum and maximum values. Conclusion: The effect of education and industrial accidents are reduced only when construction workers are motivated to participate.

• Journal of Intelligence and Information Systems
• /
• v.29 no.1
• /
• pp.107-119
• /
• 2023

The technology that recognizes a soldier's motion and movement status has recently attracted large attention as a combination of wearable technology and artificial intelligence, which is expected to upend the paradigm of troop management. The accuracy of state determination should be maintained at a high-end level to make sure of the expected vital functions both in a training situation; an evaluation and solution provision for each individual's motion, and in a combat situation; overall enhancement in managing troops. However, when input data is given as a timer series or sequence, existing feedforward networks would show overt limitations in maximizing classification performance. Since human behavior data (3-axis accelerations and 3-axis angular velocities) handled for military motion recognition requires the process of analyzing its time-dependent characteristics, this study proposes a high-performance data-driven classifier which utilizes the long-short term memory to identify the order dependence of acquired data, learning to classify eight representative military operations (Sitting, Standing, Walking, Running, Ascending, Descending, Low Crawl, and High Crawl). Since the accuracy is highly dependent on a network's learning conditions and variables, manual adjustment may neither be cost-effective nor guarantee optimal results during learning. Therefore, in this study, we optimized hyperparameters using Bayesian optimization for maximized generalization performance. As a result, the final architecture could reduce the error rate by 62.56% compared to the existing network with a similar number of learnable parameters, with the final accuracy of 98.39% for various military operations.

• Journal of the Society of Disaster Information
• /
• v.19 no.3
• /
• pp.519-531
• /
• 2023

With the aim of preventing safety accidents at construction sites, the company aims to create safe behaviors intended through variables called smart safety and health activities to help reduce industrial accidents. Purpose: It analyzes how smart safety and health activities affect accidents caused by unsafe behavior and changes in worker behavior, which is the root cause, and verifies the hypothesis that it helps prevent safety accidents and protect workers' lives. Method: Smart safety and health activities were selected as independent variables (X), and intended safety and anxiety, which are workers' behavioral intentions, were set as dependent variables (Y), attitude and subjective norms, and planned behavioral control as parameters (M). Exploratory factor analysis, discriminant validity analysis, and intensive validity analysis of safety and health activities were used to analyze the scale's reliability and validity. To verify the hypothesis of behavior change, the study was verified through Bayesian model analysis and MC simulation's probability density distribution. Result: It was found that workers who experienced smart safety and health activities at construction sites had the highest analysis of reducing unstable behavior and performing intended safety behavior. The research hypothesis that this will affect changes in worker behavior has been proven, the correlation between variables has been verified in the structural equation and path analysis of the research analysis, and it has been confirmed that smart safety and health activities can control and reduce worker instability. Conclusion: Smart safety and health activities are a very important item to prevent accidents and change workers' behavior at construction sites.

• Korean Journal of Ecology and Environment
• /
• v.56 no.2
• /
• pp.140-150
• /
• 2023

The Hangang River^* is necessary to manage the water environment of severe pollution due to the high density of residential areas, parks, and agriculture and the large population concentrated there. Benthic macroinvertebrates, such as chironomids larvae, are bioindicator species that reflect environmental changes and are crucial for water quality monitoring. In this study, we investigated morphological characteristics and molecular analysis of the chironomids larvae inhabiting the Hangang River area for water environment surveys. For this research, 20 rivers, lakes, and urban area in the Hangang River basin were selected. Chironomids larvae were collected from July to September 2022, and their appearance and characteristics were identified through morphological identification. In addition, phylogenetic analysis was performed based on the mtCOI gene sequences of the collected chironomids larvae, and identification at the genus level was confirmed. As a result, 32 species and 18 genera of 3 subfamilies of Chironomidae larvae were identified, and Stictochironomus sp. dominated most sites(6 sites). The morphological characteristics of the identified chironomids larvae, such as the mentum, ventromental plate, and antenna, were organized into table and pictorial keys, and a Bayesian inference molecular phylogeny was presented. These results provide basic morphological information for genus-level identification and can be used as fundamental information for water quality management.

• Knowledge Management Research
• /
• v.25 no.2
• /
• pp.115-134
• /
• 2024

In an age where consumers can easily search and pass on their opinions of products and purchasing decisions through the internet, Electronic-word-of-mouth(Ewom) plays an important role in decision making of other potential customers. In this study, we empirically analyze the impact EWOM on consumer purchase decisions, when contradictory Ewom is presented from varying sources of information, such as experts and general consumers. First, we find that when there is a consensus among different information sources there exists a positive relationship between Ewom sentiment and purchase intent, confirming the results of previous literature. However, when expert opinion and consumer opinion do not match we find that consumer opinion is more impactful on purchasing decisions compared to the expert opinion, regardless of product types. The findings of this study add insight to the current literature by examining the effect of contradictory Ewom on purchase decisions, and also to industry marketers by presenting a more efficient strategy in promoting positive Ewom for different product types.

• Journal of Ecology and Environment
• /
• v.48 no.2
• /
• pp.196-206
• /
• 2024

Background: Most of the biomass equations were developed using sample trees collected mainly from pan-tropical and tropical regions that may over- or underestimate biomass. Site-specific models would improve the accuracy of the biomass estimates and enhance the country's measurement, reporting, and verification activities. The aim of the study is to develop site-specific biomass estimation models and validate and evaluate the existing generic models developed for pan-tropical forest and newly developed allometric models. Total of 140 trees was harvested from each diameter class biomass model development. Data was analyzed using SAS procedures. All relevant statistical tests (normality, multicollinearity, and heteroscedasticity) were performed. Data was transformed to logarithmic functions and multiple linear regression techniques were used to develop model to estimate aboveground biomass (AGB). The root mean square error (RMSE) was used for measuring model bias, precision, and accuracy. The coefficient of determination (R² and adjusted [adj]-R²), the Akaike Information Criterion (AIC) and the Schwarz Bayesian information Criterion was employed to select most appropriate models. Results: For the general total AGB models, adj-R² ranged from 0.71 to 0.85, and model 9 with diameter at stump height at 10 cm (DSH₁₀), ρ and crown width (CW) as predictor variables, performed best according to RMSE and AIC. For the merchantable stem models, adj-R² varied from 0.73 to 0.82, and model 8) with combination of ρ, diameter at breast height and height (H), CW and DSH₁₀ as predictor variables, was best in terms of RMSE and AIC. The results showed that a best-fit model for above-ground biomass of tree components was developed. AGB_Stem = exp {-1.8296 + 0.4814 natural logarithm (Ln) (ρD²H) + 0.1751 Ln (CW) + 0.4059 Ln (DSH₃₀)} AGB_Branch = exp {-131.6 + 15.0013 Ln (ρD²H) + 13.176 Ln (CW) + 21.8506 Ln (DSH₃₀)} AGB_Foliage = exp {-0.9496 + 0.5282 Ln (DSH₃₀) + 2.3492 Ln (ρ) + 0.4286 Ln (CW)} AGB_Total = exp {-1.8245 + 1.4358 Ln (DSH₃₀) + 1.9921 Ln (ρ) + 0.6154 Ln (CW)} Conclusions: The results demonstrated that the development of local models derived from an appropriate sample of representative species can greatly improve the estimation of total AGB.

• Journal of Intelligence and Information Systems
• /
• v.21 no.1
• /
• pp.119-142
• /
• 2015

With the explosive growth in the volume of information, Internet users are experiencing considerable difficulties in obtaining necessary information online. Against this backdrop, ever-greater importance is being placed on a recommender system that provides information catered to user preferences and tastes in an attempt to address issues associated with information overload. To this end, a number of techniques have been proposed, including content-based filtering (CBF), demographic filtering (DF) and collaborative filtering (CF). Among them, CBF and DF require external information and thus cannot be applied to a variety of domains. CF, on the other hand, is widely used since it is relatively free from the domain constraint. The CF technique is broadly classified into memory-based CF, model-based CF and hybrid CF. Model-based CF addresses the drawbacks of CF by considering the Bayesian model, clustering model or dependency network model. This filtering technique not only improves the sparsity and scalability issues but also boosts predictive performance. However, it involves expensive model-building and results in a tradeoff between performance and scalability. Such tradeoff is attributed to reduced coverage, which is a type of sparsity issues. In addition, expensive model-building may lead to performance instability since changes in the domain environment cannot be immediately incorporated into the model due to high costs involved. Cumulative changes in the domain environment that have failed to be reflected eventually undermine system performance. This study incorporates the Markov model of transition probabilities and the concept of fuzzy clustering with CBCF to propose predictive clustering-based CF (PCCF) that solves the issues of reduced coverage and of unstable performance. The method improves performance instability by tracking the changes in user preferences and bridging the gap between the static model and dynamic users. Furthermore, the issue of reduced coverage also improves by expanding the coverage based on transition probabilities and clustering probabilities. The proposed method consists of four processes. First, user preferences are normalized in preference clustering. Second, changes in user preferences are detected from review score entries during preference transition detection. Third, user propensities are normalized using patterns of changes (propensities) in user preferences in propensity clustering. Lastly, the preference prediction model is developed to predict user preferences for items during preference prediction. The proposed method has been validated by testing the robustness of performance instability and scalability-performance tradeoff. The initial test compared and analyzed the performance of individual recommender systems each enabled by IBCF, CBCF, ICFEC and PCCF under an environment where data sparsity had been minimized. The following test adjusted the optimal number of clusters in CBCF, ICFEC and PCCF for a comparative analysis of subsequent changes in the system performance. The test results revealed that the suggested method produced insignificant improvement in performance in comparison with the existing techniques. In addition, it failed to achieve significant improvement in the standard deviation that indicates the degree of data fluctuation. Notwithstanding, it resulted in marked improvement over the existing techniques in terms of range that indicates the level of performance fluctuation. The level of performance fluctuation before and after the model generation improved by 51.31% in the initial test. Then in the following test, there has been 36.05% improvement in the level of performance fluctuation driven by the changes in the number of clusters. This signifies that the proposed method, despite the slight performance improvement, clearly offers better performance stability compared to the existing techniques. Further research on this study will be directed toward enhancing the recommendation performance that failed to demonstrate significant improvement over the existing techniques. The future research will consider the introduction of a high-dimensional parameter-free clustering algorithm or deep learning-based model in order to improve performance in recommendations.

• Journal of Intelligence and Information Systems
• /
• v.27 no.3
• /
• pp.175-197
• /
• 2021

Recently, with the development of deep learning technology, research on unstructured data analysis is being actively conducted, and it is showing remarkable results in various fields such as classification, summary, and generation. Among various text analysis fields, text classification is the most widely used technology in academia and industry. Text classification includes binary class classification with one label among two classes, multi-class classification with one label among several classes, and multi-label classification with multiple labels among several classes. In particular, multi-label classification requires a different training method from binary class classification and multi-class classification because of the characteristic of having multiple labels. In addition, since the number of labels to be predicted increases as the number of labels and classes increases, there is a limitation in that performance improvement is difficult due to an increase in prediction difficulty. To overcome these limitations, (i) compressing the initially given high-dimensional label space into a low-dimensional latent label space, (ii) after performing training to predict the compressed label, (iii) restoring the predicted label to the high-dimensional original label space, research on label embedding is being actively conducted. Typical label embedding techniques include Principal Label Space Transformation (PLST), Multi-Label Classification via Boolean Matrix Decomposition (MLC-BMaD), and Bayesian Multi-Label Compressed Sensing (BML-CS). However, since these techniques consider only the linear relationship between labels or compress the labels by random transformation, it is difficult to understand the non-linear relationship between labels, so there is a limitation in that it is not possible to create a latent label space sufficiently containing the information of the original label. Recently, there have been increasing attempts to improve performance by applying deep learning technology to label embedding. Label embedding using an autoencoder, a deep learning model that is effective for data compression and restoration, is representative. However, the traditional autoencoder-based label embedding has a limitation in that a large amount of information loss occurs when compressing a high-dimensional label space having a myriad of classes into a low-dimensional latent label space. This can be found in the gradient loss problem that occurs in the backpropagation process of learning. To solve this problem, skip connection was devised, and by adding the input of the layer to the output to prevent gradient loss during backpropagation, efficient learning is possible even when the layer is deep. Skip connection is mainly used for image feature extraction in convolutional neural networks, but studies using skip connection in autoencoder or label embedding process are still lacking. Therefore, in this study, we propose an autoencoder-based label embedding methodology in which skip connections are added to each of the encoder and decoder to form a low-dimensional latent label space that reflects the information of the high-dimensional label space well. In addition, the proposed methodology was applied to actual paper keywords to derive the high-dimensional keyword label space and the low-dimensional latent label space. Using this, we conducted an experiment to predict the compressed keyword vector existing in the latent label space from the paper abstract and to evaluate the multi-label classification by restoring the predicted keyword vector back to the original label space. As a result, the accuracy, precision, recall, and F1 score used as performance indicators showed far superior performance in multi-label classification based on the proposed methodology compared to traditional multi-label classification methods. This can be seen that the low-dimensional latent label space derived through the proposed methodology well reflected the information of the high-dimensional label space, which ultimately led to the improvement of the performance of the multi-label classification itself. In addition, the utility of the proposed methodology was identified by comparing the performance of the proposed methodology according to the domain characteristics and the number of dimensions of the latent label space.