• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 1996.04a
• /
• pp.76-78
• /
• 1996

When data is classified and each class has weight, the mean of data is a weighted average. When the class values and weights are trapezoidal fuzzy numbers, we can prove the weghted average is a fuzzy number though not trapezoidal. Its 4 corner points are obtained.

• Journal of the Korean Data and Information Science Society
• /
• v.16 no.4
• /
• pp.815-821
• /
• 2005

Among the various artificial neural networks the backpropagation network (BPN) has become a standard one. One of the components in a neural network is an activating function or a transfer function of which a representative function is a sigmoid. We have discovered that by updating the slope parameter of a sigmoid function simultaneous with the weights could improve performance of a BPN.

• Journal of the Korean Data and Information Science Society
• /
• v.16 no.4
• /
• pp.1045-1052
• /
• 2005

This paper deals with the estimations of kernel ridge regression when the responses are subject to randomly right censoring. The weighted data are formed by redistributing the weights of the censored data to the uncensored data. Then kernel ridge regression can be taken up with the weighted data. The hyperparameters of model which affect the performance of the proposed procedure are selected by a generalized approximate cross validation(GACV) function. Experimental results are then presented which indicate the performance of the proposed procedure.

• Asian-Australasian Journal of Animal Sciences
• /
• v.12 no.1
• /
• pp.5-8
• /
• 1999

Weight records from birth to calving and calving scores of 407 two-year old heifers and weights of their offspring from birth to one year of age were used to study the effects of sire birth weight on maternal traits of their female progeny. The heifers ($G_1$) were Ihe progeny of 81 sires ($G_0$) and were classified into three classes based on their sires' birth weights (High, Medium and Low). The heifers were from three distinct breed-groups and were mated to bulls with medium birth weights within each breed-group to produce the second generation ($G_2$). The data were analyzed using a covariance model. The female progeny of high birth-weight sires were heavier from birth to calving than those sired by medium and low birth-weight bulls. The effect of sire birth weight on calving difficulty scores of their female progeny was not significant. Grand progeny (G2) of low birth-weight sires were lighter at birth than those from high birth-weight sires (p < 0.05) but they did not differ significantly in weaning and yearling weights from the other two Grand progeny groups. The results indicated that using low birth weight sires would not result in an increase in the incidence of dystocia among their female progeny calving at two-year of age and would not have an adverse effect On weaning and yearling weights of their grand progeny.

• Korean Journal of Poultry Science
• /
• v.25 no.1
• /
• pp.11-19
• /
• 1998

This study was conducted to estimate the heritabilities and genetic correlations on egg weight and egg compositional traits for breeding plan and selection in Korean native chicken. Data analyzed were the records of 46,908 eggs from 430 layers produced from 180 dam and 26 sire families, from April, 1994 to September, 1995. On egg weight and egg compositional traits at 1st egg, 300 and 500 days of age, the egg weights were 41.489, 49.544 and 52.770g ; the albumin weights were 25.953, 29.979 and 31.288g; the yolk weights were 11.091, 14.541 and 16.368g; shell weights were 4.472, 5.037 and 5.099g, respectively. The estimates of heritability of egg weights and egg compositional traits based on the variance of sires, dams and combined components at 300 days of age were 0.214, 0.226 and 0.720 for egg weight ; 0.307, 0.152 and 0.730 for albumin weight ; 0.124, 0.953 and 0.699 for yolk weight ; 0.047, 0.026 and 0.536 for shell weight. The genetic correlation coefficient between egg weight and albumen weight was 0.083~0.951 ; 0.310~0.507 between egg weight and yolk weight ; 0.242~0.523 between egg weight and shell weight ; 0. 237~0. 413 between albumen weight and yolk weight ; 0.232~0.449 between albumen weight and shell weight ; -0.264~0.239 between yolk weight and shell weight, respectively.

• Asian-Australasian Journal of Animal Sciences
• /
• v.9 no.5
• /
• pp.509-513
• /
• 1996

The data analysed consisted of body weight records at birth, and at 3, 6, 9 and 12 months of age of 549 half Friesian $\times$ Bunaji crossbred heifers collected over a twenty-five year period (1965-1989). Least squares $means{\pm}s.e$. of body weights at birth, 3, 6, 9 and 12 months of age were $26.7{\pm}1.3$, $72.4{\pm}4.5$, $112.9{\pm}6.9$, $147.2{\pm}9.2$ and $182.1{\pm}11.1kg$, respectively. Year of birth was highly significant (p < 0.01) in affecting body weights at all ages, while the effect of month of birth was not. Seasonal influence on birth weight and body weights at 3 and 6 months of age was significant (p < 0.05). Phenotypic correlations between body weights at all ages were positive and highly significant (p < 0.01), ranging from 0.30 to 0.79. The results of this study showed that the beneficial effect of crossbreeding Friesian with Bunaji cattle was reflected in the growth performance of the $F_1$ crosses, since they grew faster than the indigenous Bunaji from brith to yearling age. The study also indicated that heifer selection for yearling body weight can be done early on the basis of weights at 3 and 6 months of age.

• Proceedings of the Korean Geotechical Society Conference
• /
• 2002.10a
• /
• pp.499-506
• /
• 2002

The purpose of this study is to develop landslide susceptibility analysis techniques using artificial neural networks and to apply the newly developed techniques for assessment of landslide susceptibility to the study area of Yongin in Korea. Landslide locations were identified in the study area from interpretation of aerial Photographs and field survey data, and a spatial database of the topography, soil type and timber cover were constructed. The landslide-related factors such as topographic slope, topographic curvature, soil texture, soil drainage, soil effective thickness, timber age, and timber diameter were extracted from the spatial database. Using those factors, landslide susceptibility and weights of each factor were analyzed by two artificial neural network methods. In the first method, the landslide susceptibility index was calculated by the back propagation method, which is a type of artificial neural network method. Then, the susceptibility map was made with a GIS program. The results of the landslide susceptibility analysis were verified using landslide location data. The verification results show satisfactory agreement between the susceptibility index and existing landslide location data. In the second method, weights of each factor were determinated. The weights, relative importance of each factor, were calculated using importance-free characteristics method of artificial neural networks.

• Journal of the Korea Convergence Society
• /
• v.12 no.7
• /
• pp.53-60
• /
• 2021

LOD-based recommender systems usually leverage the data available within LOD datasets, such as DBpedia, in order to recommend items(movies, books, music) to the end users. These systems use a semantic similarity algorithm that calculates the degree of matching between pairs of Linked Data resources. In this paper, we proposed a new approach to measuring semantic distance in an LOD-based recommender system by assigning weights converted from user ratings to links in the LOD graph. The semantic distance measurement model proposed in this paper is based on a processing step in which a graph is personalized to a user through weight calculation and a method of applying these weights to LDSD. The Experimental results showed that the proposed method showed higher accuracy compared to other similar methods, and it contributed to the improvement of similarity by expanding the range of semantic distance measurement of the recommender system. As future work, we aim to analyze the impact on the model using different methods of LOD-based similarity measurement.

• Smart Structures and Systems
• /
• v.23 no.3
• /
• pp.263-278
• /
• 2019

Moving train load parameters, including train speed, axle spacing, gross train weight and axle weights, are identified based on strain-monitoring data. In this paper, according to influence line theory, the classic moving force identification method is enhanced to handle time-varying velocity of the train. First, the moments that the axles move through a set of fixed points are identified from a series of pulses extracted from the second derivative of the structural strain response. Subsequently, the train speed and axle spacing are identified. In addition, based on the fact that the integral area of the structural strain response is a constant under a unit force at a unit speed, the gross train weight can be obtained from the integral area of the measured strain response. Meanwhile, the corrected second derivative peak values, in which the effect of time-varying velocity is eliminated, are selected to distribute the gross train weight. Hence the axle weights could be identified. Afterwards, numerical simulations are employed to verify the proposed method and investigate the effect of the sampling frequency on the identification accuracy. Eventually, the method is verified using the real-time strain data of a continuous steel truss railway bridge. Results show that train speed, axle spacing and gross train weight can be accurately identified in the time domain. However, only the approximate values of the axle weights could be obtained with the updated method. The identified results can provide reliable reference for determining fatigue deterioration and predicting the remaining service life of railway bridges.

• Journal of the Korean Data and Information Science Society
• /
• v.15 no.2
• /
• pp.331-337
• /
• 2004

In the analysis of massive volume data, a neural network model is a useful tool. To implement the Neural network model, it is important to select initial value. Since the initial values are generally used as random value in the neural network, the convergent performance and the prediction rate of model are not stable. To overcome the drawback a possible method use samples randomly selected from the whole data set. That is, coefficients estimated by logistic regression based on the samples are the initial values.