• Journal of Intelligence and Information Systems
• /
• v.29 no.1
• /
• pp.249-263
• /
• 2023

Collaborative Filtering, a representative recommendation system methodology, consists of two approaches: neighbor methods and latent factor models. Among these, the latent factor model using matrix factorization decomposes the user-item interaction matrix into two lower-dimensional rectangular matrices, predicting the item's rating through the product of these matrices. Due to the factor vectors inferred from rating patterns capturing user and item characteristics, this method is superior in scalability, accuracy, and flexibility compared to neighbor-based methods. However, it has a fundamental drawback: the need to reflect the diversity of preferences of different individuals for items with no ratings. This limitation leads to repetitive and inaccurate recommendations. The Adaptive Deep Latent Factor Model (ADLFM) was developed to address this issue. This model adaptively learns the preferences for each item by using the item description, which provides a detailed summary and explanation of the item. ADLFM takes in item description as input, calculates latent vectors of the user and item, and presents a method that can reflect personal diversity using an attention score. However, due to the requirement of a dataset that includes item descriptions, the domain that can apply ADLFM is limited, resulting in generalization limitations. This study proposes a Generalized Adaptive Deep Latent Factor Recommendation Model, G-ADLFRM, to improve the limitations of ADLFM. Firstly, we use item ID, commonly used in recommendation systems, as input instead of the item description. Additionally, we apply improved deep learning model structures such as Self-Attention, Multi-head Attention, and Multi-Conv1D. We conducted experiments on various datasets with input and model structure changes. The results showed that when only the input was changed, MAE increased slightly compared to ADLFM due to accompanying information loss, resulting in decreased recommendation performance. However, the average learning speed per epoch significantly improved as the amount of information to be processed decreased. When both the input and the model structure were changed, the best-performing Multi-Conv1d structure showed similar performance to ADLFM, sufficiently counteracting the information loss caused by the input change. We conclude that G-ADLFRM is a new, lightweight, and generalizable model that maintains the performance of the existing ADLFM while enabling fast learning and inference.

• Proceeding of KASS Symposium
• /
• 2008.05a
• /
• pp.19-24
• /
• 2008

The thesis is for gust factor needing when calculate the wind resistance design. For the gust factor, to the membrane structural model, carry through the wind tunnel test and the static load test. Therefore, at first through the tensile test of the fabric material, designate the material of the membrane structural model. Then, to saddle, wave, arch and point four kinds of basic shape membrane structural models, carry on the wind tunnel test, determine their dynamic load and distortion on lateral direction. Finally, according to distort situation of the membrane structure in the wind tunnel test, carry on the static load experiment outside of the wind tunnel, calculate static load which corresponding with distort. According to dynamic load and the static load, figure out gust factor of these kinds of basic membrane structure.

• Korean Journal of Child Studies
• /
• v.33 no.6
• /
• pp.85-107
• /
• 2012

In this study, the Infants & Toddlers' Playfulness Scale(ITPS) was developed and its reliability and validity were examined. The participants in this study consisted of 656 parents whose children were 8 to 36 months old. Five factors were identified from the results of the exploratory factor analysis performed on the 29 items that had been derived by analyzing existing research. The Goodness of Fit Indices(GFIs) of the confirmatory factor analyses, performed on the 1st-order 5-factor model and the 2nd-order one factor model, were both satisfactory. Concurrent validity was established by the high correlations of r = .631~.808(p < .01) between each factor of the ITPS and the total score of the Children's Playfulness Scale(CPS), an existing measure for children's playfulness. The reliability of each factor, as measured by Cronbach's ${\alpha}$, ranged from .773 to .883. Consequently, the ITPS developed in this study can be seen as a reliable and valid scale that can be used by parents to measure the playfulness of their infants and toddlers in a multi-dimensional manner.

• Korean Journal of Child Studies
• /
• v.37 no.6
• /
• pp.107-117
• /
• 2016

Objective: This study examined the construct validity of K-WPPSI-IV. Factor structures of the structures of the K-WPPSI-IV full scale as well as primary index scales for two age ranges (2 years, 6 months to 3 years, 11 months; 4 years to 7 years, 7 months) were examined. Methods: Data were collected from 1,700 children aged 2 years, 6 months to 7 years, 7 months during the K-WPPSI-IV standardization. Confirmatory factor analyses were conducted using the K-WPPSI-IV subtest performances with maximum likelihood estimation using Amos 18. Results: First, the three-factor model (verbal comprehension, visual spatial, and working memory) fitted best for the younger age range. However, the five-factor model (verbal comprehension, visual spatial, fluid reasoning, working memory, and processing speed) fitted best for the older age range. Residuals suggest the presence of two nested subfactors within the verbal comprehension factor (broad/expressive and focused/simple). Second, the confirmatory factor analysis on primary index subtests identified factors that account for the intercorrelations among the reduced sets of primary index subtests. Conclusion: The findings showed that the theoretical structures of WPPSI-IV subtests were confirmed within K-WPPSI-IV.

• Structural Engineering and Mechanics
• /
• v.19 no.6
• /
• pp.639-652
• /
• 2005

In this paper, a full-scale K-joint specimen was tested to failure under cyclic combined axial and in-plane bending loads. In the fatigue test, the crack developments were monitored step by step using the alternating current potential drop (ACPD) technique. Using Paris' law, stress intensity factor, which is a fracture parameter to be frequently used by many designers to predict the integrity and residual life of tubular joints, can be obtained from experimental test results of the crack growth rate. Furthermore, a scheme of automatic mesh generation for a cracked K-joint is introduced, and numerical analysis of stress intensity factor for the K-joint specimen has then been carried out. In the finite element analysis, J-integral method is used to estimate the stress intensity factors along the crack front. The numerical stress intensity factor results have been validated through comparing them with the experimental results. The comparison shows that the proposed numerical model can produce reasonably accurate stress intensity factor values. The effects of different crack shapes on the stress intensity factors have also been investigated, and it has been found that semi-ellipse is suitable and accurate to be adopted in numerical analysis for the stress intensity factor. Therefore, the proposed model in this paper is reliable to be used for estimating the stress intensity factor values of cracked tubular K-joints for design purposes.

• 재활복지
• /
• v.21 no.1
• /
• pp.47-70
• /
• 2017

The aim of this study was to explore the effects of the epidemiological factor (depression), the behavioral factor (activities of daily living), the predisposing factor (self-efficacy), reinforcing factors (family support, professional support), and enabling factor (resource availability, accessibility) on quality of life in home-dwelling disabled people in rural areas. The conceptual model for this study was established on the basis of the PRECEDE model which was developed by Green and Kreuter. Data were drawn from the "Preliminary Investigation for Community-centered rehabilitation" conducted by a public health center located in the O province in 2011 and 186 of 190 disabled people who participated in the survey were included in the final analysis. Data were analyzed using Direct effects on quality of life arose from latent variables depicting the epidemiological factor (depression)and reinforcing factor (family support, professional support), while indirect effects arose from the behavioral factor (activities of daily living), the predisposing factor (self-efficacy), and enabling support (resource availability, accessibility). This model explained 85.5% of the variance in quality of life among rural disabled individuals. These findings may have shed some light on the necessity of including strategies to reduce depression and to strengthen supports from family and healthcare professionals when performing rehabilitation programs to improve quality of life in home-dwelling disabled people in rural areas. Furthermore, it suggested that it would be useful to develop specific strategies and tactics which might increase self-efficacy and to expand linkages between public health centers and other professional institutions such as hospitals for community-centered rehabilitation services for individuals with disabilities.

• Journal of Soil and Groundwater Environment
• /
• v.25 no.3
• /
• pp.52-64
• /
• 2020

The universal soil loss equation (USLE), a model for estimating the potential soil loss, has been used not only in research areas but also in establishing national policies in South Korea. Despite its wide applicability, USLE cannot adequately address the effect of seasonal variances. To overcome this limit, the ArcGIS-based Sediment Assessment Tool for Effective Erosion (ArcSATEEC) has been developed as an alternative model. Although the field-scale (< 100 ㎡) application of this model produced reliable estimation results, it is still challenging to validate accuracy of the model estimation because it only estimates potential soil losses, not the actual sediment yield. Therefore, in this study, a method for estimating actual soil loss based on the ArcSATEEC model was suggested. The model was applied to eight watersheds in South Korea to estimate sediment yields. Correction factor was introduced for each watershed, and the estimated sediment yield was compared with that of the estimated yield by LOAD ESTimator (LOADEST). Sediment yield estimation for all watersheds exhibited reliable results, and the validity of the proposed correction factor was confirmed, suggesting the correction factor needs to be considered in estimating actual soil loss.

• Journal of Nuclear Fuel Cycle and Waste Technology(JNFCWT)
• /
• v.2 no.1
• /
• pp.35-40
• /
• 2004

Final disposal of radioactive waste generated from Nuclear Power Plant (NPP) requires the detailed information about the characteristics and the quantities of radionuclides in waste package. Most of these radionuclides are difficult to measure and expensive to assay. Thus it is suggested to the indirect method by which the concentration of the Difficult-to-Measure (DTM) nuclide is estimated using the correlations of concentration - it is called the scaling factor - between Easy-to-Measure (Key) nuclides and DTM nuclides with the measured concentration of the Key nuclide. In general, the scaling factor is determined by the log mean average (LMA) method and the regression method. However, these methods are inadequate to apply to fission product nuclides and some activation product nuclides such as 14$^{C}$ and 90$^{Sr}$ . In this study, the artificial neural network (ANN) method is suggested to improve the conventional SF determination methods - the LMA method and the regression method. The root mean squared errors (RMSE) of the ANN models are compared with those of the conventional SF determination models for 14$^{C}$ and 90$^{Sr}$ in two parts divided by a training part and a validation part. The SF determination models are arranged in the order of RMSEs as the following order: ANN model

• Tunnel and Underground Space
• /
• v.8 no.4
• /
• pp.287-295
• /
• 1998

Limit equilibrium method is widely used for the stability analysis of soil slopes. In jointed rock slopes however, the failure of the slope is largely dependent upon the strength and deformability of the joints in the rock mass and quite often failure occurs along the joints. This paper describes the use of ubiquitous joint model for the stability analysis of the jointed rock slopes. This model is essentially an anisotropic elasto-plastic model and can simulate two sets of joint in arbitrary orientations. Validation of the developed with the factor of safety equal to unity was selected when the shape of the failure plane is assumed log spiral. Then the factor of safety of the rock slope having two perpendicular joint sets was calculated while rotating joint orientations. Rusults were compared with limit equilibrium solutions on soil slopes having equivalent soil properties when plane sliding was assumed. Developed model predicted the factor of safety of jointed rock slope in a reasonable accuracy when joint spacing is sufficiently small.

• Journal of Environmental Science International
• /
• v.12 no.6
• /
• pp.605-613
• /
• 2003

In order to make use of the protection effect against wind by the vegetation, it examined whether it should make what vegetation form and arrangement using the 2-dimensional non-hydrostatic model. When the foliage shielding factor increases, it becomes hard to take in protection effect against wind in a residential section. When it makes height of vegetation high, it becomes hard to take in protection effect against wind with height. In the comparison in the case where vegetation high is gradually made low toward wind-stream from a vegetation, and the case of making it low gradually, although former tends to receive the protection effect against wind by the vegetation, attenuation of wind velicity becomes large. In the comparison in the case where foliage shielding factor and distribution of density of leaf are gathered gradually toward wind-stream from a vegetation. It has been understood to evaluate to height the influence that the vegetation multi-layer model by which the heat revenue and expenditure in the direction of the vegetation height is considered is used, and to characterize the vegetation group by the parameter setting.