Generalization and Selection of Examples in Feedforward Neural Networks

2000 ◽  
Vol 12 (10) ◽  
pp. 2405-2426 ◽  
Author(s):  
Leonardo Franco ◽  
Sergio A. Cannas
Keyword(s):  
Analytical Calculation ◽  
Feedforward Neural Networks ◽  
Global Learning ◽  
Parity Problem ◽  
Learning Procedure ◽  
Minimum Number ◽  
Training Examples ◽  
General Architecture ◽  
Independent Criterion ◽  
Selection Of

In this work, we study how the selection of examples affects the learning procedure in a boolean neural network and its relationship with the complexity of the function under study and its architecture. We analyze the generalization capacity for different target functions with particular architectures through an analytical calculation of the minimum number of examples needed to obtain full generalization (i.e., zero generalization error). The analysis of the training sets associated with such parameter leads us to propose a general architecture-independent criterion for selection of training examples. The criterion was checked through numerical simulations for various particular target functions with particular architectures, as well as for random target functions in a nonoverlapping receptive field perceptron. In all cases, the selection sampling criterion lead to an improvement in the generalization capacity compared with a pure random sampling. We also show that for the parity problem, one of the most used problems for testing learning algorithms, only the use of the whole set of examples ensures global learning in a depth two architecture. We show that this difficulty can be overcome by considering a tree-structured network of depth 2 log2(N) – 1.

Determination of the minimum number of technological bases of a part

2020 ◽  
pp. 73-75
Author(s):  
B.M. Bazrov ◽  
T.M. Gaynutdinov
Keyword(s):  
Cost Price ◽  
Labor Input ◽  
Part Surface ◽  
Technological Base ◽  
Size Accuracy ◽  
Minimum Number ◽  
The Cost ◽  
Input Cost ◽  
Selection Of

The selection of technological bases is considered before the choice of the type of billet and the development of the route of the technological process. A technique is proposed for selecting the minimum number of sets of technological bases according to the criterion of equality in the cost price of manufacturing the part according to the principle of unity and combination of bases at this stage. Keywords: part, surface, coordinating size, accuracy, design and technological base, labor input, cost price. [email protected]

Observational Effects on the Food Preferences of Children With Autism Spectrum Disorder

2020 ◽  
pp. 108835762095436
Author(s):  
Joseph H. Cihon ◽  
Mary Jane Weiss ◽  
Julia L. Ferguson ◽  
Justin B. Leaf ◽  
Thomas Zane ◽  
...  
Keyword(s):  
Autism Spectrum Disorder ◽  
Observational Learning ◽  
Food Preferences ◽  
Autism Spectrum ◽  
Food Selectivity ◽  
Spectrum Disorder ◽  
Simple Task ◽  
Learning Procedure ◽  
Reversal Design ◽  
Selection Of

Research addressing food selectivity has involved intrusive techniques such as escape extinction. It is possible that observational learning methods employed in previous studies could provide the desired results with respect to food selectivity without the need for invasive physical interventions. The purpose of this study was to evaluate the effectiveness of an observational learning procedure on the selection of food items of three children diagnosed with autism spectrum disorder. Baseline consisted of a simple task after which a choice was presented between high- and low-preferred foods. The intervention consisted of observing an adult engage in the same task and selecting the low-preferred food while making favorable comments and engaging with the food in novel ways. The results of a reversal design demonstrated that selection of the low-preferred food only occurred following the introduction of the intervention, and all three participants engaged in flexible responding as a result of the intervention.

scholarly journals TERRESTRIAL LASER SCANNERS SELF-CALIBRATION STUDY: DATUM CONSTRAINTS ANALYSES FOR NETWORK CONFIGURATIONS

2015 ◽  
Vol XL-2/W4 ◽  
pp. 1-9 ◽  
Author(s):  
M. A. Abbas ◽  
H. Setan ◽  
Z. Majid ◽  
A. K. Chong ◽  
L. Chong Luh ◽  
...  
Keyword(s):  
Laser Scanner ◽  
Network Configuration ◽  
Calibration Technique ◽  
Self Calibration ◽  
Laser Scanners ◽  
Different Types ◽  
Minimum Number ◽  
Calibration Study ◽  
Different Sources ◽  
Selection Of

Similar to other electronic instruments, terrestrial laser scanner (TLS) can also inherent with various systematic errors coming from different sources. Self-calibration technique is a method available to investigate these errors for TLS which were adopted from photogrammetry technique. According to the photogrammetry principle, the selection of datum constraints can cause different types of parameter correlations. However, the network configuration applied by TLS and photogrammetry calibrations are quite different, thus, this study has investigated the significant of photogrammetry datum constraints principle in TLS self-calibration. To ensure that the assessment is thorough, the datum constraints analyses were carried out using three variant network configurations: 1) minimum number of scan stations; 2) minimum number of surfaces for targets distribution; and 3) minimum number of point targets. Based on graphical and statistical, the analyses of datum constraints selection indicated that the parameter correlations obtained are significantly similar. In addition, the analysis has demonstrated that network configuration is a very crucial factor to reduce the correlation between the calculated parameters.

NEUROSLAB – neural network system for horizontal formwork selection

1995 ◽  
Vol 22 (4) ◽  
pp. 785-792 ◽  
Author(s):  
Awad S. Hanna ◽  
Ahmed B. Senouci
Keyword(s):  
Neural Network ◽  
Expert System ◽  
Back Propagation ◽  
Network System ◽  
Neural Network System ◽  
Testing Error ◽  
Training Error ◽  
Training Examples ◽  
High Level ◽  
Selection Of

This paper presents an overview of the neural network technique as a tool for concrete formwork selection. The paper discusses the development and the implementation of a neural network system, NEUROSLAB, for the selection of horizontal formwork systems. A rule-based expert system for the selection of horizontal systems, SLABFORM, was used as the basis for the development of NEUROSLAB. A training set of 202 cases was used to train the network. The network adequately learned the training examples with an average training error of 0.025. A set of 50 cases was used to test the generalization ability of the system. The network was able to accurately select the appropriate horizontal formwork system with an average testing error of 0.057. The ability of the network to deal with noisy data was also tested. Up to 50% noise was added to the data and introduced to the network. The results showed that the network presented could accurately identify the appropriate horizontal formwork system at high level of noise. Finally, the solution chosen by an expert was compared to that produced by the network. The network was able to mimic the expert's formwork selection. Key words: formwork, horizontal formwork systems, neural network, formwork selection, back propagation, expert system.

scholarly journals Inductive analysis of the deflection of a beam truss allowing kinematic variation

2018 ◽  
Vol 239 ◽  
pp. 01012
Author(s):  
Mikhail Kirsanov ◽  
Evgeny Komerzan ◽  
Olesya Sviridenko
Keyword(s):  
Optimal Design ◽  
External Load ◽  
Analytical Calculation ◽  
System Of Equations ◽  
Number Of Solutions ◽  
Recurrence Equations ◽  
General Terms ◽  
Cutting Out ◽  
Selection Of ◽  
Inductive Analysis

A scheme of a statically determinate planar truss is proposed and an analytical calculation of its deflection and displacement of the mobile support are obtained. The forces in the rods from the external load, uniformly distributed over the nodes of the lower or upper belt, are determined by the method of cutting out nodes using the computer mathematic system Maple. In the generalization of a number of solutions of trusses with a different number of panels to the general case, the general terms of the sequence of coefficients in the formulas are found from solutions of linear homogeneous recurrence equations. To compose and solve these equations, Maple operators were used. In the process of calculation it was revealed that for even numbers of panels in half the span, the determinant of the system of equations degenerates. This corresponds to the kinematic degeneracy of the structure. The corresponding scheme of possible speeds of the truss is given. The displacement was determined by the Maxwell-Mohr’s formula. The graphs of the obtained dependences have appreciable jumps, which in principle can be used in the selection of optimal design sizes.

Optimal Selection of Long Time Acoustic Features Using GA for the Assessment of Vocal Fold Disorders

2012 ◽  
Vol 239-240 ◽  
pp. 65-70 ◽  
Author(s):  
Sindhu Ravindran ◽  
Neoh Siew-Chin ◽  
Hariharan Muthusamy
Keyword(s):  
Vocal Fold ◽  
Acoustic Analysis ◽  
Feature Reduction ◽  
Optimal Selection ◽  
Acoustical Parameters ◽  
Long Time ◽  
Minimum Number ◽  
Reduction Methods ◽  
The Voice ◽  
Selection Of

In recent times, vocal fold problems have been increasing dramatically due to unhealthy social habits and voice abuse. Non-invasive methods like acoustic analysis of voice signals can be used to investigate such problems. Various feature extraction techniques are used to classify the voice signals into normal and pathological. Among them, long-time acoustical parameters are used by many researchers. The selection of best long-time acoustical parameters is very important to reduce the computational complexity, as well as to achieve better accuracy with minimum number of features. In order to select best long-time acoustical parameters, different feature reduction methods or feature selection methods are proposed by researchers. In this work, genetic algorithm (GA) based optimal selection of long-time acoustical parameters is proposed to achieve higher accuracy with minimum number of features. The classification is carried out using k-nearest neighbourhood (k-NN) classifier. In comparison with other works in the literature, the simulation results show that a minimum of 5 features are required to classify the voice signals by GA and a better accuracy of 94.29% is achieved.

scholarly journals Optimization of Binder for Improving Strength and Shatter Index of Briquettes for BOF Dust using Design of Experiments

2019 ◽  
Vol 9 (1) ◽  
pp. 6282-6287
Keyword(s):  
Low Temperature ◽  
Design Of Experiments ◽  
Room Temperature ◽  
Waste Utilization ◽  
Taguchi Technique ◽  
Plant Waste ◽  
Minimum Number ◽  
Cold Bonding ◽  
Bonding Technique ◽  
Selection Of

Effectiveness of Recycling of steel plant waste is very much dependent on agglomeration technique. Sintering, pelletization and briquetting are some of the techniques which are frequently used for waste utilization. Aim of this study is to prepare composite briquettes by cold bonding technique, by which phsico-chemical changesoccurred at room temperature or low temperature. Two binders are mixed in proportion to achieve the required properties specifically strength and shatter index. The design of experiments is used to find the proper combination of binders to get the optimum value of properties. Experimental work for the same is carried out in such a way that minimum number of experiment can give output as desired. For this ‘Design of Experiment’ methodology is applied to select the runs of experiment. After the selection of orthogonal array and experiment combinations, Taguchi technique is used with two variable (starch and molasses) and three levels (2.5%, 5% and 7.5% of each) i.e. L9 Array to analyze the results. Minitab15 software is used. Conclusion and comments are based on the same.

scholarly journals Repeatability estimates and minimum number of evaluations for selection of elephant-grass genotypes for herbage production

2020 ◽  
Vol 36 (1) ◽  
Author(s):  
Erina Vitório Rodrigues ◽  
Rogério Figueiredo Daher ◽  
Geraldo de Amaral Gravina ◽  
Alexandre Pio Viana ◽  
Maria do Socorro Bezerra de Araújo ◽  
...  
Keyword(s):  
Plant Height ◽  
Principal Components ◽  
Dry Matter ◽  
Block Design ◽  
Phenotypic Selection ◽  
Stem Diameter ◽  
Dry Matter Yield ◽  
Elephant Grass ◽  
Minimum Number ◽  
Selection Of

In forage-plants breeding, the selection of superior genotypes has been undertaken through successive harvests in previously established intervals. However, this process involves many steps, the evaluation of many traits, and a great spending with costs and labor. Thus the estimate of the repeatability is essential in improvement of perennials, it allows predicting genotypic value of the individual, the minimum number of evaluations in the selection of genotypes and minimizes resources and time in the selection of promising individuals. The objective of this study was to estimate the repeatability coefficient for morphological traits in elephant grass and determine the number of evaluations needed for phenotypic selection more efficient. The experimental randomized block design with 53 genotypes and two replications. The repeatability coefficients were estimated for variables plant height, number of tillers, stem diameter and dry matter yield, using the methods of Anova, Principal Components and Structural Analysis. We observed significant differences between genotypes (P <0.01) for all variables. The main components provide larger estimates of repeatability when compared to other methods. Estimates of the repeatability coefficients are of high magnitude average for the variables plant height (0.44) number of tillers (0.44) and stem diameter (0.63) and low magnitude for dry matter production (0.27). The Principal Components method requires five, five, two and eleven measurements for plant height, number of tillers, stem diameter and dry matter yield, respectively, with 80% reliability.  

Sign in / Sign up

Export Citation Format

Share Document