Search

There are several problems for the school children to take the pure milk especially the goat milk every day because of the unpleasant taste and smell of the goat milk are not so good. Besides, the pure milk has a short life time. It is difficult to store the goat milk in a long period of time at room temperature. The solution for this problem is change the goat milk from liquid form into the chewable milk tablet that can be easily consume by the school children. Goat milk in the form of tablet formulation will increase the shelf life of the product. Therefore, the chewable goat milk tablet can be stored in a long period of time, convenience, and suitable for distribution to remote areas. In this project, the focusing was more into the optimization of the weight adjustment of the tableting machine before and after the dry granulation to form good chewable tablet. The pre-evaluation had been done for the goat milk powder and granule after dry granulation process. It can be concluded that goat milk tablet can be an alternative to the pure goat milk. The poorly flowing properties of goat milk powder can be improved by using dry granulation techniques. Based on this study, F4D shows the best setting to produce the goat milk tablet using dry granulation technique. F4D successfully improved the flow ability and compressibility of goat milk tablet. It also produce the best goat milk tablet as it possess uniformity of weight, low friability and good hardness

To estimate the nutritional requirements of hair sheep, knowledge about the animal’s weight and its relationships with growth performances is essential. A study was carried with the objective to establish the relationships between BW, fasting BW (FBW), empty BW (EBW), average daily gain (ADG) and empty BW gain (EBWG) for hair sheep in growing and finishing phases in Brazilian conditions. Databases were obtained from 32 studies, for a total of 1145 observations; there were 3 sex classes (non-castrated male, castrated male and female) and 2 feeding systems (pasture and feedlot). The most representative breeds in the database were Santa Ines (n = 473), Morada Nova (n = 70) and Brazilian Somali (n = 47). The other animals in the database were crossbreeds (n = 555). The FBW (kg), EBW and EBWG (kg/day) were estimated according to linear regression. A random coefficient model was adopted, considering the study as a random effect and including the possibility of covariance between the slope and the intercept. The coefficients obtained from the linear regression of the FBW against the BW, EBW against the FBW and EBWG against the ADG did not differ between sex class (P > 0.05) and genotype (P > 0.05). The equations generated to estimate FBW from the BW, EBW from the FBW and EBWG from the ADG are as follows: FBW = −0.5470 (±0.2025) + 0.9313(±0.019) × BW, EBW = −1.4944 (±0.3639) + 0.8816 (±0.018) × FBW and EBWG = 0.906 (±0.019) × ADG, respectively. The low mean squared error values found in the cross-validation confirmed the reliability of these equations. Considering a sheep with a BW of 30 kg and a 100 g ADG, the estimated FBW, EBW and EBWG calculated using the generated equations are 27, 22.65 and 0.090 kg, respectively. In conclusion, the generated equations can be used in growing hair sheep. The validation procedure applied to the generated equations showed that its use for hair sheep seems to be appropriate.

The objective of this study is to reduce the coffee mixed powder weight loss during filling into sticks by volumetric filling machine. Fishbone Diagram based on 4M analysis (i.e., Man, Machine, Material and Method) was applied to analyze the root causes and find solutions to reduce coffee mixed powder weight loss. This study was conducted as an action research in the practical situation by studying machine function of volumetric filling machine in the processing line and observing practices of operators. From brainstorming, the major root causes impacted coffee mixed powder weight loss were machine conditions with improper machine’s basic condition setting and weight adjustment by operators. The actions and experiments were executed by restoring basic conditions of volumetric filling machine and modifying some parts of equipment, e.g., design of agitator in hopper and design of weight scale. Visual controls and inspection standard for operators were also developed for control purposes. The improvement was carried out for three months consecutively. The powder weight loss from filling machine was successfully reduced from 0.38% to 0.24%, which was 36.84% improved.

This study aims to identify the type of coffee powder aroma from the coffee beans blending using backpropagation artificial neural network (ANN). Backpropagation is a controlled training implementing a weight adjustment pattern to achieve a minimum error value between the the predicted and the actual output. Within this study, the coffee aroma testing utilized electronic tasting sensor system consisted of 4 sensors namely TGS 2611, TGS 2620, TGS 2610 and TGS 2602. The coffee aroma monitoring and data collection in this system applied LabVIEW software as a virtual instrumentation. The testing result of this ANN was able to distinguish the coffee variety of Robusta, Arabica coffee powder and the one without any coffee aroma. The backpropagation architecture was formed by 3 layers consisting of 1 input layer with 4 input nerve cells, 1 hidden layer with 8 neural cells, and 2 output layers by applying the backpropagation training algorithm. The training data was taken from 70 data samples of each circumstance of coffee with 5 testing times. The results of the training and testing showed that the established backpropagation was capable to identify and differenciate the coffee powder in accordance with the given input with different average success rate; 91.96% for Robusta coffee, 100 % for Arabica coffee, and no 84.24% for without coffee aroma.

Aiming at the problems of obstacle avoidance and bullet avoidance during the patrol swarm, this paper analyzed the defects of the classical artificial potential field, proposed an adjustable escape method, which establish the relationship between the adjustment coefficient and the distance. This method avoid too large or too small escape force that get the bullet into new local shock problem near the target. Then given the weight calculation and parameter selection method, restricted the escape motion by kinematics according to the constraints in the actual motion. This improved method can effecting solve the problem of avoidance in dynamic and complex environment.

We proposed an improved weighted SVM for imbalanced classification.Different factor scores are computed by categorizing instances based on the SVM margin.Numerical experiments showed the proposed method outperforming the existing approaches in terms of F-measure AUC. To address class imbalance in data, we propose a new weight adjustment factor that is applied to a weighted support vector machine (SVM) as a weak learner of the AdaBoost algorithm. Different factor scores are computed by categorizing instances based on the SVM margin and are assigned to related instances. The SVM margin is used to define borderline and noisy instances, and the factor scores are assigned to only borderline instances and positive noise. The adjustment factor is then employed as a multiplier to the instance weight in the AdaBoost algorithm when learning a weighted SVM. Using 10 real class-imbalanced datasets, we compare the proposed method to a standard SVM and other SVMs combined with various sampling and boosting methods. Numerical experiments show that the proposed method outperforms existing approaches in terms of F-measure and area under the receiver operating characteristic curve, which means that the proposed method is useful for relaxing the class-imbalance problem by addressing well-known degradation issues such as overlap, small disjunct, and data shift problems.

The krill herd (KH) algorithm is an emerging meta-heuristic algorithm for solving complex problems. Though it is robust in optimization, there are some parameters that need to be fine-tuned for improved performance. This paper proposes an entropy-based inertia weight krill herd (EBIWKH) algorithm, which could adaptively adjust the inertia weight according to the variance of population entropy. It is tested on CEC2017 benchmark functions and compared with other inertia weight adjustment strategies. Experimental results show that the EBIWKH algorithm is more robust and stable than other compared algorithms.

Use of Murthy’s method in estimation of population parameters, such as population totals, population means, and population variances has been limited to surveys where survey data values are complete. This study applies weight adjustment technique to estimate a population total under simple random sampling without replacement. The asymptotic properties show that the estimated population total is sufficient for the true population total. The proposed estimator is obtained by symmetrizing Murthy’s estimator.

Imputation is a commonly used method to handle missing survey data. The performance of the imputation method is influenced by various factors, especially an outlier. The removal of the outlier in a data set is a simple and effective approach to reduce the effect of an outlier. In this paper in order to improve the precision of multiple imputation, we study a imputation method which reduces the effect of outlier using various weight adjustment methods that include the removal of an outlier method. The regression method in PROC/MI in SAS is used for multiple imputation and the obtained final adjusted weight is used as a weight variable to obtain the imputed values. Simulation studies compared the performance of various weight adjustment methods and Monthly Labor Statistic data is used for real data analysis.

Fractional regression hot deck imputation (FRHDI) imputes multiple values for each instance of a missing dependent variable. The imputed values are equal to the predicted value plus multiple random residuals. Fractional weights enable variance estimation and preserve correlations. In some circumstances with some starting weight values, existing procedures for computing FRHDI weights can produce negative values. We discuss procedures for constructing non-negative adjusted fractional weights for FRHDI and study performance of the algorithm using simulation. The algorithm can be used effectively with FRDHI procedures for handling missing data in the context of a complex sample survey.

This research compares the two features within an images, namely, color and edge characteristic and adjusting weight ratios from color weight value 1, edge weight value 1, and 3 color to edge weight ratios; 0.7 to 0.3, 0.5 to 0.5, and 0.3 to 0.7 to retrieve the most relevance query outputs. Based on the results, Color to edge weight ratio at 0.5 to 0.5 retrieved the most relevance query images outputs on RGB color models. The recall value and accuracy remained robust while F-measurements is fair. The HSV model evaluation was fair on both recall value and accuracy, but F-measure signified revision.

In order to solve the problem how to determine the quantitative relationship of characteristic factors caused by uncertainties, this paper describes the target case by characteristic factors in case-based reasoning, determines the initial weight according to the characteristic factor matrix, introduces Hebb learning rule to make a weight adjustment and applies the improved Euclidean distance to calculate the similarity between the cases to look for the most similar case. The similarity comparison between CBR based on Grey Correlation and the proposed method proves this model a better performance. At last, the application example of the prediction model after an earthquake can prove the feasibility.

The predictive ability of Artificial Neural Network (ANN) on the effect of the concentration (30, 40, 50 y 60 % w/w) and temperature (30, 40 y 50°C) of fructooligosaccharides solution, in the mass, moisture, volume and solids of osmodehydrated yacon cubes, and in the coefficients of the water means effective diffusivity with and without shrinkage was evaluated. The Feedforward type ANN with the Backpropagation training algorithms and the Levenberg-Marquardt weight adjustment was applied, using the following topology: 10-5 goal error, 0.01 learning rate, 0.5 moment coefficient, 2 input neurons, 6 output neurons, one hidden layer with 18 neurons, 15 training stages and logsig-pureline transfer functions. The overall average error achieved by the ANN was 3.44% and correlation coefficients were bigger than 0.9. No significant differences were found between the experimental values and the predicted values achieved by the ANN and with the predicted values achieved by a statistical model of second-order polynomial regression (p > 0.95). Se evaluó la capacidad predictiva de la Red Neuronal Artificial (RNA) en el efecto de la concentración (30, 40, 50 y 60 % p/p) y temperatura (30, 40 y 50°C) de la solución de fructooligosacaridos (FOS) en la masa, humedad, volumen y sólidos en cubos de yacón osmodeshidratados, y en el coeficiente de difusividad efectiva media del agua, con y sin encogimiento. Se aplicó la RNA del tipo Feedforward con los algoritmos de entrenamiento Backpropagation y de ajuste de pesos Levenberg-Marquardt, usando la topología: error meta de 10-5, tasa de aprendizaje de 0.01, coeficiente de momento de 0.5, 2 neuronas de entrada, 6 neuronas de salida, una capa oculta con 18 neuronas, 15 etapas de entrenamiento y funciones de transferencia logsigpurelin. El error promedio global por la RNA fue 3.44% y los coeficientes de correlación fueron mayores a 0.9. No se encontraron diferencias significativas entre los valores experimentales con los valores predichos por la RNA y con los valores predichos por un modelo estadístico de regresión polinomial de segundo orden (p > 0.95).

BLS weight adjustment is a widely used method for business surveys with non-responses and outliers. Recent surveys show that the non-response weight adjustment of the BLS method is the same as the ratio imputation method. In this paper, we suggested a modified BLS weight adjustment method by imputing missing values instead of using weight adjustment for non-response. Monthly labor survey data is used for a small Monte-Carlo simulation and we conclude that the suggested method is superior to the original BLS weight adjustment method.

Weighting is a common form of unit nonresponse adjustment in sample surveys where entire questionnaires are missing due to noncontact or refusal to participate. A common approach computes the response weight as the inverse of the response rate within adjustment cells based on covariate information. In this paper, we consider the efficiency and robustness of nonresponse weight adjustment bated on the response propensity and predictive mean. In the simulation study based on 2000 Fishry Census in Korea, the root mean squared errors for assessing the various ways of forming nonresponse adjustment cell s are investigated. The simulation result suggest that the most important feature of variables for inclusion in weighting adjustment is that they are predictive of survey outcomes. Though useful, prediction of the propensity to response is a secondary. Also the result suggest that adjustment cells based on joint classification by the response propensity and predictor of the outcomes is productive.

BLS adjustment methods have been able to provide more accurate estimates of total and make samples represent population characteristics by post-adjustment of design weights of samples. However, BLS methods use additional data, for instance number of employee, without this information or using other information, give different weight adjustment factors. In this paper we studied the sensitivity of the variables used in BLS adjustment. The 2007 monthly labor survey data is used in analysis.

This note comments on a paper by Triantaphyllou (J. Multi-Crit. Decis. Anal. 2001; 10: 11–25) that attempts to demonstrate new types of rank reversal that can occur with the analytic hierarchy process (AHP). He contends that the reversals are attributable to the various types of normalization that are used with the addition step in AHP synthesis. His paper goes on to suggest that the multiplicative AHP should be used instead. This note shows that the cause of the problem is another one: AHP’s independence axiom, which prohibits adjusting the criteria weights when the set of alternatives or the type of normalization change. If the criteria weights are adjusted properly, none of the rank reversals will occur

A non-discriminating criterion is defined as a criterion where the decision-maker is indifferent among the alternatives. One would therefore expect the final rank order of the alternatives not to be affected by removing it. A previously published paper by Finan and Hurley (Comput. Oper. Res. 2002; 29: 1025–1030) showed that in the analytic hierarchy process removing such a criterion from a multilevel hierarchy can reverse rank. In this paper, we offer an explanation of this particular rank reversal phenomenon and show how it can be avoided. We do this by taking into account that there is a link between the normalization and weighting processes, which suggests adjusting appropriate weights when removing criteria. Further, we discuss whether a non-discriminating criterion should be removed in the first place

In sample design, it is usually planned by purpose and the range of the announcing statistics from the survey. After survey, getting a proper and decent statistics, applying the proper weights on the results of survey is very important and necessary. Therefore in this study, three estimation methods which are raking, BLS and general linear regression method are compared with MSE, Coverage, CV, LE and NC.

The augmented-neural-network (AugNN) approach has been applied lately to some NP-Hard combinatorial problems, such as task scheduling, open-shop scheduling and resource-constraint project scheduling. In this approach the problem of search in the solution-space is transformed to a search in a weight-matrix space, much like in a neural-network approach. Some weight adjustment strategies are then used to converge to a good set of weights for a locally optimal solution. While empirical results have demonstrated the effectiveness of the AugNN approach vis-a-vis a few other metaheuristics, little theoretical insights exist which justify this approach and explain the effectiveness thereof. This paper provides some theoretical insights and justification for the AugNN approach through some basic theorems and also describes the algorithm and the formulation with the help of examples.

In sample design, determining the weights of estimates becomes usually great influence on the result. In this article, raking methods are applied to different domain and depending on the range of the domain and sample size, the results of estimates are explained and compared. For the comparison, we use the MSE, MAE, MSPE and MAPE with Actual State of Minor Enterprisers Human Resources Survey data in 2001. The simulation result shows that more elaborate method is superior to the widely used method as expected but the difference is not quite significant.

Weighted coefficients of image retrieval algorithm based on relevance feedback are determined in advance, which is lack of flexibility. In order to obtain satisfactory retrieval results, this algorithm requires a large amount of feedback calculation and efficiency of the algorithm is low. Aiming at the faults of relevance feedback, the adaptive adjustment algorithm of weighted coefficients based on quantum particle swarm optimization is presented, which is composed of user feedback process and particle evolution process. The particle encoding process and fitness function calculation process are worked out. The result of experiment using the Corel standard library, shows that quantum particle swarm optimization algorithm greatly improves the retrieval accuracy than the other image retrieval algorithms.

This paper studies weight adjustment in multiple attributes group decision making model. Firstly, the attribute weights respectively calculated by methods of AHP and entropy are optimized based on objective programming model. Secondly, the initial weights of decision-makers are updated by the grey correlation degree between the individual decision results and the group decision results. Furtherly, the group decision results are updated. Then the weight-adjusting is continued on the basis of the new group decision results. Based on 2-Norm Minkowski measure, the steady weights of decision-makers and final results of the group are gotten with the process of adjustment. Finally, a numerical example to evaluate wind power equipment supplier shows the feasibility and practicability of the proposed algorithm.

Particle Swarm Optimization (PSO) is a popular optimization technique which is inspired by the social behavior of birds flocking or fishes schooling for finding food. It is a new metaheuristic search algorithm developed by Eberhart and Kennedy in 1995. However, the standard PSO has a shortcoming, i.e., premature convergence and easy to get stack or fall into local optimum. Inertia weight is an important parameter in PSO, which significantly affect the performance of PSO. There are many variations of inertia weight strategies have been proposed in order to overcome the shortcoming. In this paper, a new modified PSO with random activation to increase exploration ability, help trapped particles for jumping-out from local optimum and avoid premature convergence is proposed. In the proposed method, an inertia weight is decreased linearly until half of iteration, and then a random number for an inertia weight is applied until the end of iteration. To emphasis the role of this new inertia weight adjustment, the modified PSO paradigm is named Modified PSO with random activation (MPSO-RA). The experiments with three famous benchmark functions show that the accuracy and success rate of the proposed MPSO-RA increase of 43.23% and 32.95% compared with the standard PSO.

Serial ultrasonic measures of fat thickness (FTU) and longissimus muscle area (LMU) were taken on 180 feedlot steers representing 11 sire-breed groups at two 60-d intervals. Cattle then were slaughtered in four groups of 45 at 21-d intervals with ultrasonic measures taken at each interval. Analyses of variance indicated weight (WT ) effects ( P < .001) for all FTU measures and many of the LMU measures. Age was a source of variation ( P < .05) for FTU and LMU over the first five measurement dates. Sire-breed (SBD) effects (P < .001) were prevalent for all FTU estimates and early LMU predictions. Some dam line and SBD x dam line interactions were detected (P < .1) for the FTU and LMU estimates. Age effects were described quadratically (P < .001) when FTU and LMU were regressed on age (R 2 = 47 and 68%, respectively). When FTU and LMU were regressed on WT, effects were described quadratically ( P < .001) with R 2 values of 46 and 72%, respectively. Animals then were assigned to one of four biological types ( BT ) based on percentage of retail product (0-mm fat trim) using discriminant cluster analysis. Within BT, ultrasonic variables were regressed on age. Quadratic effects (P < .01) were detected for both FTU and LMU. Weight also contributed to variation observed in both ultrasonically measured variables for all BT ( P < .001). These data indicate that changes in fat and muscle size as affected by age and weight are detectable using serially collected ultrasound measures. Furthermore, the age and weight relationships of predicted FTA and LMA varied across the four biological types. This suggests a need for derivation of biological type-specific age and (or) weight adjustment equations for cattle evaluation purposes.

Presents a comparative study of different methods for realizing the basic learning algorithm of dynamic fuzzy neural networks (DFNNs). Performances of the least squared estimation, Kalman filter and extended Kalman filter methods used for weight adjustment in the basic learning algorithm of DFNNs in terms of learning speed, neuron requirement, approximation accuracy and noise immunity are evaluated and compared.

We analyze different methods of controlling magnification factor in the relation between the density of the input vectors and the asymptotic density of the quantizers. We also propose a new method of modeling the one-dimensional self-organizing process and examine some properties of the method. Empirical studies show strong influence of the learning rate strategies on the observed magnification factor of the vector quantization process.

This paper presents a method of automatically adjusting aberrational weighting factors in a damped least squares optimization program, based on tolerance data provided for each aberration site. The method which is based on previous work by the authors and that of Yao has been developed into a tolerance-based method for aberrational weight adjustment. This enhanced method allows the user to set and adjust weights by a new tolerance-based method. The use of a power index ((rho) ) in the tolerance-based method provides continuously variable control over the amount by which the aberrational weights are adjusted. The practical use of the tolerance-based method is demonstrated using simple and well understood test lenses to reveal the modifications that take place within the iterative design progress, that result from aberrational weight adjustment. Suggestions for suitable tolerance schemes have also been investigated as an integral part of this work, based on the work of Hopkins.

Applications of simulated annealing to the optimization of radiation treatment plans, in which a set of beam weights are iteratively adjusted so as to minimize a cost function, have been motivated by its potential for finding the global or near‐global minimum among multiple minima. However, the method has been found to be slow, requiring several tens of thousands of iterations to optimize 50 to 100 variables. A technique to improve the efficiency for finding a solution is reported, which is generally applicable to the optimization of continuous variables. In previous applications of simulated annealing to treatment planningoptimization, only one or two weights are varied each iteration. This approach is to change all weights simultaneously, using random changes that are initially large to coarsely sample the cost function, then are reduced with iteration to probe finer structure. The performance of different methods are compared in optimizing a plan for treatment of the prostate, in which the search space consists of 54 noncoplanar beams and the cost function is based on tumor control and normal tissue complication probabilities. The proposed method yields solutions with similar values of the cost function in only a fraction of the iterations compared either to a fixed single weight adjustment technique, or to a method which combines the Nelder and Mead downhill simplex with simulated annealing.