scholarly journals SVM: Object Detection using Rotation-Invariant and Fisher Discriminative Convolutional Neural Networks

2021 ◽  
pp. 36-39
Author(s):  
V. Pavithra ◽  
R. Kannan
Keyword(s):  
Neural Networks ◽  
Test Results ◽  
Rotation Invariant ◽  
Improvement Strategy ◽  
Characterization Result ◽  
Proposed Model ◽  
Vector Machines ◽  
High Measurement ◽  
Face Datasets ◽  
Solid System

With the touchy development of information, the multi-see information is broadly utilized in numerous fields, for example, information mining, AI, PC vision, etc. Since such information consistently has a perplexing construction, for example numerous classes, numerous viewpoints of portrayal and high measurement, how to detail a precise and solid system for the multi-see order is an exceptionally difficult assignment. In this paper, we propose a novel multi-see characterization technique by utilizing various multi-class Backing Vector Machines (SVMs) with a novel shared system. Here each multi-class SVM installs the scaling component to renewedly change the weight assignment, everything being equal, which is useful to feature more significant and discriminative highlights. Besides, we receive the choice capacity esteems to incorporate different multi-class students and present the certainty score across various classes to decide the last characterization result. Moreover, through a progression of the numerical derivation, we connect the proposed model with the reasonable issue and address it through a rotating emphasis improvement strategy. We assess the proposed strategy on a few picture and face datasets, and the test results exhibit that our proposed technique performs better compared to other cutting edge learning calculations.

Application of Chaos Particle Swarm Optimized Neural Networks for Evaluating the Credit Risk

2011 ◽  
Vol 460-461 ◽  
pp. 687-691
Author(s):  
Zhi Bin Xiong
Keyword(s):  
Neural Networks ◽  
Credit Risk ◽  
Simple Structure ◽  
Particle Swarm ◽  
Pso Algorithm ◽  
Test Results ◽  
Chaos Optimization ◽  
Proposed Model ◽  
The Neural Networks ◽  
Better Than

This paper proposes a hybrid algorithm based on chaos optimization and particle swarm optimization (PSO) to improve the performance of the neural networks (NN) on evaluating credit risk. The hybrid algorthm not only maintains the advantage of simple structure, but also improves the convergence of the traditional PSO algorithm, and enhances the global optimization capability and accuracy of the algorithm. The test results indicate that the performance of the proposed model is better than the ones of NN model using the BP algorithm and traditional PSO algorithm.

scholarly journals RECURRENT NEURAL NETWORKS AND NONLINEAR PREDICTION IN SUPPORT VECTOR MACHINES

2019 ◽  
Vol 2019 (1) ◽  
pp. 33-40 ◽  
Author(s):  
Jennifer S. Raj ◽  
Vijitha Ananthi J
Keyword(s):  
Neural Networks ◽  
Support Vector Machines ◽  
Recurrent Neural Networks ◽  
Learning Algorithm ◽  
Back Propagation ◽  
Support Vector ◽  
Nonlinear Prediction ◽  
Proposed Model ◽  
Vector Machines ◽  
Robust Prediction

The nonlinear regression estimation issues are solved by successful application of a novel neural network technique termed as support vector machines (SVMs). Evaluation of recurrent neural networks (RNNs) can assist in pattern recognition of several real-time applications and reduce the pattern mismatch. This paper provides a robust prediction model for multiple applications. Traditionally, back-propagation algorithms were used for training RNN. This paper predict system reliability by applying SVM learning algorithm to RNN. Comparison of the proposed model is done with the existing systems for analysis of prediction performance. These results indicate that the performance of proposed system exceeds that of the existing ones.

In-Plane Flexible Ring Tire Model—Part 1: Modeling and Parameter Identification

2018 ◽  
Vol 46 (3) ◽  
pp. 174-219 ◽  
Author(s):  
Bin Li ◽  
Xiaobo Yang ◽  
James Yang ◽  
Yunqing Zhang ◽  
Zeyu Ma
Keyword(s):  
Static Load ◽  
Model Parameters ◽  
Tire Model ◽  
Test Results ◽  
Lumped Mass ◽  
Virtual Test ◽  
Proposed Model ◽  
Particle Swarm Method ◽  
Vehicle Dynamic Simulation ◽  
Flexible Ring

ABSTRACT The tire model is essential for accurate and efficient vehicle dynamic simulation. In this article, an in-plane flexible ring tire model is proposed, in which the tire is composed of a rigid rim, a number of discretized lumped mass belt points, and numerous massless tread blocks attached on the belt. One set of tire model parameters is identified by approaching the predicted results with ADAMS® FTire virtual test results for one particular cleat test through the particle swarm method using MATLAB®. Based on the identified parameters, the tire model is further validated by comparing the predicted results with FTire for the static load-deflection tests and other cleat tests. Finally, several important aspects regarding the proposed model are discussed.

scholarly journals Neural Network Approach for Global Solar Irradiance Prediction at Extremely Short-Time-Intervals Using Particle Swarm Optimization Algorithm

Energies ◽  
2021 ◽  
Vol 14 (4) ◽  
pp. 1213
Author(s):  
Ahmed Aljanad ◽  
Nadia M. L. Tan ◽  
Vassilios G. Agelidis ◽  
Hussain Shareef
Keyword(s):  
Neural Networks ◽  
Particle Swarm Optimization ◽  
Solar Irradiance ◽  
Particle Swarm ◽  
Time Interval ◽  
Time Intervals ◽  
Swarm Optimization ◽  
Backpropagation Neural Networks ◽  
Proposed Model ◽  
Short Time

Hourly global solar irradiance (GSR) data are required for sizing, planning, and modeling of solar photovoltaic farms. However, operating and controlling such farms exposed to varying environmental conditions, such as fast passing clouds, necessitates GSR data to be available for very short time intervals. Classical backpropagation neural networks do not perform satisfactorily when predicting parameters within short intervals. This paper proposes a hybrid backpropagation neural networks based on particle swarm optimization. The particle swarm algorithm is used as an optimization algorithm within the backpropagation neural networks to optimize the number of hidden layers and neurons used and its learning rate. The proposed model can be used as a reliable model in predicting changes in the solar irradiance during short time interval in tropical regions such as Malaysia and other regions. Actual global solar irradiance data of 5-s and 1-min intervals, recorded by weather stations, are applied to train and test the proposed algorithm. Moreover, to ensure the adaptability and robustness of the proposed technique, two different cases are evaluated using 1-day and 3-days profiles, for two different time intervals of 1-min and 5-s each. A set of statistical error indices have been introduced to evaluate the performance of the proposed algorithm. From the results obtained, the 3-days profile’s performance evaluation of the BPNN-PSO are 1.7078 of RMSE, 0.7537 of MAE, 0.0292 of MSE, and 31.4348 of MAPE (%), at 5-s time interval, where the obtained results of 1-min interval are 0.6566 of RMSE, 0.2754 of MAE, 0.0043 of MSE, and 1.4732 of MAPE (%). The results revealed that proposed model outperformed the standalone backpropagation neural networks method in predicting global solar irradiance values for extremely short-time intervals. In addition to that, the proposed model exhibited high level of predictability compared to other existing models.

Modeling of cutting parameters in turning of PEEK composite using artificial neural networks and adaptive-neural fuzzy inference systems

2021 ◽  
pp. 089270572110130
Author(s):  
Gökçe Özden ◽  
Mustafa Özgür Öteyaka ◽  
Francisco Mata Cabrera
Keyword(s):  
Neural Networks ◽  
Artificial Neural Networks ◽  
Cutting Forces ◽  
Fuzzy Inference ◽  
Test Results ◽  
Inference System ◽  
Good Prediction Accuracy ◽  
Peek Composite ◽  
Artificial Neural ◽  
Neural Fuzzy

Polyetheretherketone (PEEK) and its composites are commonly used in the industry. Materials with PEEK are widely used in aeronautical, automotive, mechanical, medical, robotic and biomechanical applications due to superior properties, such as high-temperature work, better chemical resistance, lightweight, good absorbance of energy and high strength. To enhance the tribological and mechanical properties of unreinforced PEEK, short fibers are added to the matrix. In this study, Artificial Neural Networks (ANNs) and the Adaptive-Neural Fuzzy Inference System (ANFIS) are employed to predict the cutting forces during the machining operation of unreinforced and reinforced PEEK with30 v/v% carbon fiber and 30 v/v% glass fiber machining. The cutting speed, feed rate, material type, and cutting tools are defined as input parameters, and the cutting force is defined as the system output. The experimental results and test results that are predicted using the ANN and ANFIS models are compared in terms of the coefficient of determination ( R2) and mean absolute percentage error. The test results reveal that the ANFIS and ANN models provide good prediction accuracy and are convenient for predicting the cutting forces in the turning operation of PEEK.

A Method of Environmental Sound Classification Based on Residual Networks and Data Augmentation

2021 ◽  
pp. 2150018
Author(s):  
Jinfang Zeng ◽  
Youming Li ◽  
Yu Zhang ◽  
Da Chen
Keyword(s):  
Neural Networks ◽  
Data Augmentation ◽  
Machine Learning Techniques ◽  
Deep Convolutional Neural Networks ◽  
Mel Frequency Cepstral Coefficients ◽  
Environmental Sound ◽  
Sound Classification ◽  
Learning Techniques ◽  
Proposed Model ◽  
Audio Data

Environmental sound classification (ESC) is a challenging problem due to the complexity of sounds. To date, a variety of signal processing and machine learning techniques have been applied to ESC task, including matrix factorization, dictionary learning, wavelet filterbanks and deep neural networks. It is observed that features extracted from deeper networks tend to achieve higher performance than those extracted from shallow networks. However, in ESC task, only the deep convolutional neural networks (CNNs) which contain several layers are used and the residual networks are ignored, which lead to degradation in the performance. Meanwhile, a possible explanation for the limited exploration of CNNs and the difficulty to improve on simpler models is the relative scarcity of labeled data for ESC. In this paper, a residual network called EnvResNet for the ESC task is proposed. In addition, we propose to use audio data augmentation to overcome the problem of data scarcity. The experiments will be performed on the ESC-50 database. Combined with data augmentation, the proposed model outperforms baseline implementations relying on mel-frequency cepstral coefficients and achieves results comparable to other state-of-the-art approaches in terms of classification accuracy.

Sign in / Sign up

Export Citation Format

Share Document