• 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.

• Steel and Composite Structures
• /
• v.37 no.6
• /
• pp.711-731
• /
• 2020

This paper deals with free vibration of FG sandwich annular sector plates on Pasternak elastic foundation with different boundary conditions, based on the three-dimensional theory of elasticity. The plates with simply supported radial edges and arbitrary boundary conditions on their circular edges are considered. The influence of carbon nanotubes (CNTs) waviness, aspect ratio, internal pores and graphene platelets (GPLs) on the vibrational behavior of functionally graded nanocomposite sandwich plates is investigated in this research work. The distributions of CNTs are considered functionally graded (FG) or uniform along the thickness of upper and bottom layers of the sandwich sectorial plates and their mechanical properties are estimated by an extended rule of mixture. In this study, the classical theory concerning the mechanical efficiency of a matrix embedding finite length fibers has been modified by introducing the tube-to-tube random contact, which explicitly accounts for the progressive reduction of the tubes' effective aspect ratio as the filler content increases. The core of structure is porous and the internal pores and graphene platelets (GPLs) are distributed in the matrix of core either uniformly or non-uniformly according to three different patterns. The elastic properties of the nanocomposite are obtained by employing Halpin-Tsai micromechanics model. A semi-analytic approach composed of 2D-Generalized Differential Quadrature Method (2D-GDQM) and series solution is adopted to solve the equations of motion. The fast rate of convergence and accuracy of the method are investigated through the different solved examples. Some new results for the natural frequencies of the plate are prepared, which include the effects of elastic coefficients of foundation, boundary conditions, material and geometrical parameters. The new results can be used as benchmark solutions for future researches.

• Interaction and multiscale mechanics
• /
• v.2 no.3
• /
• pp.321-332
• /
• 2009

Recent development in composites containing phase-transforming particles, such as vanadium dioxide or barium titanate, reveals the overall stiffness and viscoelastic damping of the composites may be unbounded (Lakes et al. 2001, Jaglinski et al. 2007). Negative stiffness is induced from phase transformation predicted by the Landau phase transformation theory. Although this unbounded phenomenon is theoretically supported with the composite homogenization theory, detailed stress analyses of the composites are still lacking. In this work, we analyze the stress distribution of the Hashin-Shtrikman (HS) composite and its two-dimensional variant, namely a circular inclusion in a square plate, under the assumption that the Young's modulus of the inclusion is negative. Assumption of negative stiffness is a priori in the present analysis. For stress analysis, a closed form solution for the HS model and finite element solutions for the 2D composite are presented. A static loading condition is adopted to estimate the effective modulus of the composites by the ratio of stress to average strain on the loading edges. It is found that the interfacial stresses between the circular inclusion and matrix increase dramatically when the negative stiffness is so tuned that overall stiffness is unbounded. Furthermore, it is found that stress distributions in the inclusion are not uniform, contrary to Eshelby's theorem, which states, for two-phase, infinite composites, the inclusion's stress distribution is uniform when the shape of the inclusion has higher symmetry than an ellipse. The stability of the composites is discussed from the viewpoint of deterioration of perfect interface conditions due to excessive interfacial stresses.

• Structural Engineering and Mechanics
• /
• v.75 no.4
• /
• pp.425-434
• /
• 2020

Practical non-synoptic fluctuating wind often exhibits nonstationary features and should be modeled as nonstationary random processes. Generally, the coherence function of the fluctuating wind field has time-varying characteristics. Some studies have shown that there is a big difference between the fluctuating wind field of the coherent function model with and without time variability. Therefore, it is of significance to simulate nonstationary fluctuating wind field with time-varying coherent function. However, current studies on the numerical simulation of nonstationary fluctuating wind field with time-varying coherence are very limited, and the proposed approaches are usually based on the traditional spectral representation method with low simulation efficiency. Especially, for the simulation of multi-variable wind field of large span structures such as transmission tower-line, not only the simulation is inefficient but also the matrix decomposition may have singularity problem. In this paper, it is proposed to conduct the numerical simulation of nonstationary fluctuating wind field in one-spatial dimension with time-varying coherence based on the wavenumber-frequency spectrum. The simulated multivariable nonstationary wind field with time-varying coherence is transformed into one-dimensional nonstationary random waves in the simulated spatial domain, and the simulation by wavenumber frequency spectrum is derived. So, the proposed simulation method can avoid the complicated Cholesky decomposition. Then, the proper orthogonal decomposition is employed to decompose the time-space dependent evolutionary power spectral density and the Fourier transform of time-varying coherent function, simultaneously, so that the two-dimensional Fast Fourier transform can be applied to further improve the simulation efficiency. Finally, the proposed method is applied to simulate the longitudinal nonstationary fluctuating wind velocity field along the transmission line to illustrate its performances.

• Geophysics and Geophysical Exploration
• /
• v.9 no.1
• /
• pp.86-92
• /
• 2006

The full-waveform inversion algorithm using normalised seismic wavefields can avoid potential inversion errors due to source estimation required in conventional full-waveform inversion methods. In this paper, we have modified the inversion scheme to install a weighted smoothness constraint for better resolution, and to implement a staged approach using normalised wavefields in order of increasing frequency instead of inverting all frequency components simultaneously. The newly developed scheme is verified by using a simple two-dimensional fault model. One of the most significant improvements is based on introducing weights in model parameters, which can be derived from integrated sensitivities. The model-parameter weighting matrix is effective in selectively relaxing the smoothness constraint and in reducing artefacts in the reconstructed image. Simultaneous multiple-frequency inversion can almost be replicated by multiple single-frequency inversions. In particular, consecutively ordered single-frequency inversion, in which lower frequencies are used first, is useful for computation efficiency.

• Composites Research
• /
• v.33 no.1
• /
• pp.30-37
• /
• 2020

Stress distributions dependent on fiber arrangements are studied using the two-dimensional representative volume element (RVE) model for uni-directionally continuous fiber-reinforced composites subjected to transverse tensile loading. It is easily expected that the stresses around the fibers are concentrated mainly due to the stiffness mismatch between the fiber and matrix materials. In this presentation, it is shown that the stresses are not always increased although the distance between two fibers is shortened. The 2D RVE models, originally having a regular hexagonal fiber array, is utilized to study the effect of the fiber locations on the stress distributions. As the central fiber is relocated, the stress distributions around the fiber are obtained through finite element analysis. It is found that the stresses around the fiber are strongly dependent on the fiber distance as well as the angle between the loading direction and the line connecting two fibers.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.25 no.2
• /
• pp.139-150
• /
• 2019

Though, fishing vessel accidents account for 70 % of all maritime accidents in Korean waters, most research has focused on identifying causes and developing mitigation policies in an attempt to reduce this rate. However, predicting and evaluating accident risk needs to be done before the implementation of such reduction measures. For this reasons, we havve performed a risk analysis to calculate the risk of accidents and propose a risk criteria matrix with 4 quadrants, within one of which forecasted risk is plotted for the relative comparison of risks. For this research, we considered 9 types of fishing vessel accidents as reported by Korea Maritime Safety Tribunal (KMST). Given that no risk evaluation criteria have been established in Korea, we established a two-dimensional frequency-consequence grid consisting of four quadrants into which paired frequency and consequence for each type of accident are presented. With the simple structure of the evaluation model, one can easily verify the effect of frequency and consequence on the resulting risk within each quadrant. Consequently, these risk evaluation results will help a decision maker employ more realistic risk mitigation measures for accident types situated in different quadrants. As an application of the risk evaluation matrix, accident types were further analyzed using accident causes including human error (factor) and appropriate risk reduction options may be established by comparing the relative frequency and consequence of each accident cause.

• Proceedings of the Korean Society For Composite Materials Conference
• /
• 2002.10a
• /
• pp.88-91
• /
• 2002

The two dimensional size effect of specimen gauge section (length x width) was investigated on the compressive behavior of a T300/924 [45/-45/0/90]3s, carbon fiber-epoxy laminate. A modified ICSTM compression test fixture was used together with an anti-buckling device to test 3mm thick specimens with a 30$\times$30, 50$\times$50, 70$\times$70, and 90mm$\times$90mm gauge length by width section. In all cases failure was sudden and occurred mainly within the gauge length. Post failure examination suggests that $0^{\circ}$ fiber microbuckling is the critical damage mechanism that causes final failure. This is the matrix dominated failure mode and its triggering depends very much on initial fiber waviness. It is suggested that manufacturing process and quality may play a significant role in determining the compressive strength. When the anti-buckling device was used on specimens, it was showed that the compressive strength with the device was slightly greater than that without the device due to surface friction between the specimen and the device by pretoque in bolts of the device. In the analysis result on influence of the anti-buckling device using the finite element method, it was found that the compressive strength with the anti-buckling device by loaded bolts was about 7% higher than actual compressive strength. Additionally, compressive tests on specimen with an open hole were performed. The local stress concentration arising from the hole dominates the strength of the laminate rather than the stresses in the bulk of the material. It is observed that the remote failure stress decreases with increasing hole size and specimen width but is generally well above the value one might predict from the elastic stress concentration factor. This suggests that the material is not ideally brittle and some stress relief occurs around the hole. X-ray radiography reveals that damage in the form of fiber microbuckling and delamination initiates at the edge of the hole at approximately 80% of the failure load and extends stably under increasing load before becoming unstable at a critical length of 2-3mm (depends on specimen geometry). This damage growth and failure are analysed by a linear cohesive zone model. Using the independently measured laminate parameters of unnotched compressive strength and in-plane fracture toughness the model predicts successfully the notched strength as a function of hole size and width.

• Polymer(Korea)
• /
• v.31 no.2
• /
• pp.160-167
• /
• 2007

This paper proposes a model for the solutions predicting the coefficient of thermal expansion of composites including fiber-like shaped$(a_1>a_2=a_3)$ and disk-like shaped$(a_1=a_2>a_3)$ inclusions like two dimensional geometries, which was analyzed by one axis and a single aspect ratio, $(\rho_\alpha=a_1/a_3)$. The analysis follows the procedure developed for elastic moduli by using the Lee and Paul's approach. The effects of the aspect ratio on the coefficient of thermal expansion of composites containing aligned isotropic inclusions are examined. This model should be limited to analyze the composites with unidirectionally aligned inclusions and with complete binding to each other of both matrix and inclusions having homogeneous properties. The longitudinal coefficients of thermal expansion $\alpha_{11}$ decrease and approach the coefficient of thermal expansion of filler, as the aspect ratios increase. However, the transverse coefficients of thermal expansion $\alpha_{33}$ increase or decrease with the aspect ratios.

• Plant Biotechnology Reports
• /
• v.4 no.3
• /
• pp.201-211
• /
• 2010

We present a highly effective T-DNA inserted gene screening method as part of a reverse genetics model system using the Chinese cabbage (Brassica rapa L. spp. pekinensis). Three-step two-dimensional (2D) matrix strategies are potentially accurate and useful for the identification of specific T-DNA inserted mutants from a large population. To construct our Chinese cabbage model, we utilized a forward genetics screening approach for the abnormal phenotypes that were obtained from transgenic plants of Brassica rapa generated with Agrobacteria tumefaciens containing the pRCV2 vector. From one transgenic plant with an abnormal phenotype, we observed that the st1 gene (which is related to senescence-associated process proteins) contained a T-DNA fragment, and that its expression level was decreased. This T-DNA insert was then used as a control to construct an effective screening pool. As a result, the optimum template concentration was found to be 0.1-1 ng in our PCR strategy. For other conditions, positive changes to the Gibbs free energy prevented the formation of oligo dimers and hairpin loop structures, and autosegment extension gave better results for long fragment amplification. Using this effective reverse genetics screening method, only 23 PCR reactions were necessary to select a target gene from a pool of 100 individual DNAs. Finally, we also confirmed that the sequence we obtained from the above method was identical to the flanking sequence isolated by rescue cloning.