scholarly journals Determining Hidden Neurons with Variant Experiments in Multilayer Perception using Machine Learning Neural Networks

2019 ◽  
Vol 8 (9) ◽  
pp. 2725-2729
Keyword(s):  
Neural Network ◽  
Adaptive Learning ◽  
Intermediate Layer ◽  
Activation Function ◽  
Vital Role ◽  
Learning Ability ◽  
Output Layer ◽  
Hidden Layer ◽  
Hidden Neurons

Neural network has broadly been employed in various fields for its efficacy and its superiority. Excellence results provided can be directly provided in various analyses. Besides the variant types of neural network, Multi layer perceptron plays a vital role for its adaptive learning ability. The network makes prediction based on learn of training set. Neural has three layers then the layers are the Input, Hidden and the Output Layers. There may be more than one hidden layer but there is one input and output layer. The hidden or the intermediate layer is considered as an engine of the complete network as it has the non linear activation function and they has a sensational domination in the finishing result . The amount of neurons in three layers determines the excellence of the network. The neuron in the input and the output layer is fixed as per the dataset while for the intermediate layer it is fixed by the user in random. Increase in neuron cause over-fitting while decrease cause under fitting and these assumptions have a great impact in the final outcome. This paper discusses the existing approaches for fixing the hidden neurons and proposes a method to fix the neurons in the intermediate layer and analyse the quality of the group. The proposed procedure has variant approaches to determine the hidden neuron and they are compared. The experiment is done in WEKA and the accuracy is checked with measures.

scholarly journals Perancangan Model Jaringan Syaraf Tiruan Propagasi Balik Untuk Mengestimasi Curah Hujan

2012 ◽  
Vol 9 (2) ◽  
Author(s):  
Elohansen Padang
Keyword(s):  
Neural Network ◽  
Gradient Descent ◽  
Back Propagation ◽  
Activation Function ◽  
Back Propagation Neural Network ◽  
Rainfall Data ◽  
Training Method ◽  
Output Layer ◽  
Input Layer ◽  
Hidden Layer

This research was conducted to investigate the ability of backpropagation artificial neural network in estimating rainfall. Neural network used consists of input layer, 2 hidden layers and output layer. Input layer consists of 12 neurons that represent each input; first hidden layer consists of 12 neurons with activation function tansig, while the second hidden layer consists of 24 neurons with activation function logsig. Output layer consists of 1 neuron with activation function purelin. Training method used is the method of gradient descent with momentum. Training method used is the method of gradient descent with momentum. Learning rate and momentum parameters defined respectively by 0.1 and 0.5. To evaluate the performance of the network model to recognize patterns of rainfall data is used in Biak city rainfall data from January 1997 - December 2008 (12 years). This data is divided into 2 parts, namely training and testing data using rainfall data from January 1997-December 2005 and data estimation using rainfall data from January 2006-December 2008. From the results of this study concluded that rainfall patterns Biak town can be recognized quite well by the model of back propagation neural network. The test results and estimates of the model results testing the value of R = 0.8119, R estimate = 0.53801, MAPE test = 0.1629, and MAPE estimate = 0.6813.

scholarly journals Application of a multilayer perceptron artificial neural network for identification of peach cultivars based on physical characteristics

PeerJ ◽  
2021 ◽  
Vol 9 ◽  
pp. e11529
Author(s):  
Adel M. Al-Saif ◽  
Mahmoud Abdel-Sattar ◽  
Abdulwahed M. Aboukarima ◽  
Dalia H. Eshra
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Multilayer Perceptron ◽  
Activation Function ◽  
Cross Entropy ◽  
Output Layer ◽  
Entropy Error ◽  
Artificial Neural ◽  
The Cross ◽  
Hidden Layer

In the fresh fruit industry, identification of fruit cultivars and fruit quality is of vital importance. In the current study, nine peach cultivars (Dixon, Early Grande, Flordaprince, Flordastar, Flordaglo, Florda 834, TropicSnow, Desertred, and Swelling) were evaluated for differences in skin color, firmness, and size. Additionally, a multilayer perceptron (MLP) artificial neural network was applied for identification of the cultivars according to these attributes. The MLP was trained with an input layer including six input nodes, a single hidden layer with six hidden nodes, and an output layer with nine output nodes. A hyperbolic tangent activation function was used in the hidden layer and the cross entropy error was given because the softmax activation function was functional to the output layer. Results showed that the cross entropy error was 0.165. The peach identification process was significantly affected by the following variables in order of contribution (normalized importance): polar diameter (100%), L∗ (89.0), b∗ (88.0%), a∗ (78.5%), firmness (71.3%), and cross diameter (37.5.3%). The MLP was found to be a viable method of peach cultivar identification and classification because few identifying attributes were required and an overall classification accuracy of 100% was achieved in the testing phase. Measurements and quantitative discrimination of peach properties are provided in this research; these data may help enhance the processing efficiency and quality of processed peaches.

Experimental and Comparison Based Study on Diabetes Prediction Using Artificial Neural Network

2019 ◽  
Vol 12 ◽  
Author(s):  
Nitesh Pradhan ◽  
VijayPal Singh Dhaka ◽  
Satish Chandra Kulhari
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Nearest Neighbor ◽  
Activation Function ◽  
Support Vector ◽  
K Nearest Neighbor ◽  
Output Layer ◽  
Diabetes Prediction ◽  
Artificial Neural ◽  
Hidden Layer

Background: Diabetes is spreading in the entire world. In a survey, it is observed that every generation from child to old age people are suffering from diabetes. If diabetes is not identified in time, it may lead to deadliest disease. Prediction of diabetes is of the utmost challenging task by machines. In the human body, diabetes is one of the perilous maladies that creates depended disease such as kidney disease, heart attack, blindness etc. Thus it is very important to diagnose diabetes in time. Objective: Our target is to develop a system using Artificial Neural Network(ANN), with the ability to predict whether a patient suffers from diabetes or not. Method: This paper illustrates various machine learning techniques in form of literature review; such as Support Vector Machine, Naïve Bayes, K Nearest Neighbor, Decision Tree, Random Forest Etc. We applied ANN to predict diabetes. In this paper, the architecture of ANN consists of four hidden layers each of six neurons and one output layer with one neuron. Optimizer used for the architecture is ‘Adam’. Results: We have Pima Indian diabetes dataset of sufficient number of patients with nine different symptoms with respect to the patients and nine different features in connection with the mathematical computation/prediction. Hence we bifurcate the dataset into training and testing set in majority and minority ratio of 80:20 respectively. It facilitates us the majority patient’s data to be used as training set and minority data to be used as testing set. We train our network for multiple epoch with different activation function. We used four hidden layers with six neurons in each hidden layer and one output layer. On the hidden layer, we used multiple activation functions such as sigmoid, ReLU etc. and obtained beat accuracy (88.71%) in 600 epochs with ReLU activation function. On the output layer, we used only sigmoid activation function because we have only two classes in our dataset. Conclusion: Diabetes prediction by machine is a challenging task. So many machine learning algorithms exist to predict the diabetes such as Naïve Bayes, decision tree, K nearest neighbor, support vector machine etc. This paper presents a novel approach to predict whether a patient has diabetes or not based on Pima Indian diabetes dataset. In this paper, we used artificial neural network to train out network and it is observed that artificial neural network approach performs better than all other classifiers

scholarly journals An Intelligent Approach to Recognizing Human Movements Based on a Wearable Device

2021 ◽  
Author(s):  
Anastasya Grecheneva ◽  
Nikolay Dorofeev ◽  
Maxim Goryachev
Keyword(s):  
Neural Network ◽  
Basic Structure ◽  
Activation Function ◽  
Training Sample ◽  
Wearable Device ◽  
Output Layer ◽  
The Neural Network ◽  
Human Movements ◽  
Hidden Layer ◽  
Propagation Network

n this paper, we consider the possibility of distinguishing the movements of a person and people by their gait based on data obtained from the accelerometer of a wearable device. A mobile phone was used as a wearable device. The paper considers the features of recognizing human movements based on a wearable device. A recognition algorithm based on a neural network with preliminary data processing and correlation analysis is proposed. The volume of the training sample consisted of 32 subjects with various physiological characteristics. The sample size in the subgroup of four people ranged from 2000 to 3000 movements. The main motor patterns for classification were the movements performed when walking in a straight line and stairs with a load (a bag with a laptop weighing 3.5 kg) and without it. The direct propagation network is chosen as the basic structure for the neural network. The neural network has 260 input neurons, 100 neurons in one hidden layer, and 4 neurons in the output layer. When training the neural network, the gradient reverse descent function was used. Cross- entropy was used as an optimization criterion. The activation function of the hidden layer was a sigmoid, and the output layer was a normalized exponential function. The presented algorithm makes it possible to distinguish between subjects when performing different movements in more than 90% of cases. The practical application of the results of the work is relevant for automated information systems of the medical, law enforcement and banking sectors.

scholarly journals Influence of hinge length and distribution number on the camber and cornering properties of a non-pneumatic tire

2020 ◽  
Vol 12 (12) ◽  
pp. 168781402098468
Author(s):  
Xianbin Du ◽  
Youqun Zhao ◽  
Yijiang Ma ◽  
Hongxun Fu
Keyword(s):  
Neural Network ◽  
Network Model ◽  
Adaptive Learning ◽  
High Speed ◽  
Back Propagation ◽  
Lateral Force ◽  
Learning Ability ◽  
Vehicle Performance ◽  
Neural Network Theory ◽  
Bp Neural Network Model

The camber and cornering properties of the tire directly affect the handling stability of vehicles, especially in emergencies such as high-speed cornering and obstacle avoidance. The structural and load-bearing mode of non-pneumatic mechanical elastic (ME) wheel determine that the mechanical properties of ME wheel will change when different combinations of hinge length and distribution number are adopted. The camber and cornering properties of ME wheel with different hinge lengths and distributions were studied by combining finite element method (FEM) with neural network theory. A ME wheel back propagation (BP) neural network model was established, and the additional momentum method and adaptive learning rate method were utilized to improve BP algorithm. The learning ability and generalization ability of the network model were verified by comparing the output values with the actual input values. The camber and cornering properties of ME wheel were analyzed when the hinge length and distribution changed. The results showed the variation of lateral force and aligning torque of different wheel structures under the combined conditions, and also provided guidance for the matching of wheel and vehicle performance.

Breast cancer diagnosis using multiple activation deep neural network

2021 ◽  
pp. 1063293X2110251
Author(s):  
K Vijayakumar ◽  
Vinod J Kadam ◽  
Sudhir Kumar Sharma
Keyword(s):  
Breast Cancer ◽  
Neural Network ◽  
Deep Neural Network ◽  
Breast Cancer Diagnosis ◽  
Activation Function ◽  
Linear Functions ◽  
Cancer Data ◽  
Final Layer ◽  
Improved Performance ◽  
Hidden Layer

Deep Neural Network (DNN) stands for multilayered Neural Network (NN) that is capable of progressively learn the more abstract and composite representations of the raw features of the input data received, with no need for any feature engineering. They are advanced NNs having repetitious hidden layers between the initial input and the final layer. The working principle of such a standard deep classifier is based on a hierarchy formed by the composition of linear functions and a defined nonlinear Activation Function (AF). It remains uncertain (not clear) how the DNN classifier can function so well. But it is clear from many studies that within DNN, the AF choice has a notable impact on the kinetics of training and the success of tasks. In the past few years, different AFs have been formulated. The choice of AF is still an area of active study. Hence, in this study, a novel deep Feed forward NN model with four AFs has been proposed for breast cancer classification: hidden layer 1: Swish, hidden layer, 2:-LeakyReLU, hidden layer 3: ReLU, and final output layer: naturally Sigmoidal. The purpose of the study is twofold. Firstly, this study is a step toward a more profound understanding of DNN with layer-wise different AFs. Secondly, research is also aimed to explore better DNN-based systems to build predictive models for breast cancer data with improved accuracy. Therefore, the benchmark UCI dataset WDBC was used for the validation of the framework and evaluated using a ten-fold CV method and various performance indicators. Multiple simulations and outcomes of the experimentations have shown that the proposed solution performs in a better way than the Sigmoid, ReLU, and LeakyReLU and Swish activation DNN in terms of different parameters. This analysis contributes to producing an expert and precise clinical dataset classification method for breast cancer. Furthermore, the model also achieved improved performance compared to many established state-of-the-art algorithms/models.

scholarly journals The effect of Kurtosis on the accuracy of artificial neural network predictive model

2018 ◽  
Vol 204 ◽  
pp. 02018
Author(s):  
Aisyah Larasati ◽  
Anik Dwiastutik ◽  
Darin Ramadhanti ◽  
Aal Mahardika
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Predictive Model ◽  
Predictive Models ◽  
Output Layer ◽  
Input Layer ◽  
Artificial Neural ◽  
Misclassification Rates ◽  
Hidden Layer ◽  
Program Data

This study aims to explore the effect of kurtosis level of the data in the output layer on the accuracy of artificial neural network predictive models. The artificial neural network predictive models are comprised of one node in the output layer and six nodes in the input layer. The number of hidden layer is automatically built by the program. Data are generated using simulation approach. The results show that the kurtosis level of the node in the output layer is significantly affect the accuracy of the artificial neural network predictive model. Platycurtic and leptocurtic data has significantly higher misclassification rates than mesocurtic data. However, the misclassification rates between platycurtic and leptocurtic is not significantly different. Thus, data distribution with kurtosis nearly to zero results in a better ANN predictive model.

Prediction Model of Flow-Induced Noise in Large-Scale Centrifugal Pumps Based on BP Neural Network

2017 ◽  
Author(s):  
Chang Guo ◽  
Ming Gao ◽  
Peixin Dong ◽  
Yuetao Shi ◽  
Fengzhong Sun
Keyword(s):  
Neural Network ◽  
Network Model ◽  
Working Conditions ◽  
Bp Neural Network ◽  
Centrifugal Pumps ◽  
Output Layer ◽  
Sample Data ◽  
Input Layer ◽  
Flow Induced Noise ◽  
Hidden Layer

As one kind of serious environmental problems, flow-induced noise in centrifugal pumps pollutes the working circumstance and deteriorates the performance of pumps, meanwhile, it always changes drastically under various working conditions. Consequently, it is extremely significant to predict flow-induced noise of centrifugal pumps under various working conditions with a practical mathematical model. In this paper, a three-layer back propagation (BP) neural network model is established and the number of input, hidden and output layer node is set as 3, 6 and 1, respectively. To be specific, the flow rate, rotational speed and medium temperature are chosen as input layer, and the corresponding flow-induced noise evaluated by average of total sound pressure level (A_TSPL) as output layer. Furthermore, the tansig function is used to act as transfer function between the input layer and hidden layer, and the purelin function is used between hidden layer and output layer. The trainlm function based on Levenberg-Marquardt algorithm is selected as the training function. By using a large number of sample data, the training of the network model and prediction research are accomplished. The results indicate that good correlation is established among the sample data, and the predictive values show great consistence with simulation ones, of which the average relative error of A_TSPL in process of verification is 0.52%. The precision of the model can satisfy the requirement of relevant research and engineering application.

scholarly journals Crude Palm Oil Prediction Based on Backpropagation Neural Network Approach

2019 ◽  
Vol 2 (1) ◽  
pp. 1
Author(s):  
Hijratul Aini ◽  
Haviluddin Haviluddin
Keyword(s):  
Neural Network ◽  
Palm Oil ◽  
Good Accuracy ◽  
Activation Function ◽  
Crude Palm Oil ◽  
Backpropagation Neural Network ◽  
Neural Network Approach ◽  
Learning Function ◽  
Hidden Layer ◽  
Accuracy Of Prediction

Crude palm oil (CPO) production at PT. Perkebunan Nusantara (PTPN) XIII from January 2015 to January 2018 have been treated. This paper aims to predict CPO production using intelligent algorithms called Backpropagation Neural Network (BPNN). The accuracy of prediction algorithms have been measured by mean square error (MSE). The experiment showed that the best hidden layer architecture (HLA) is 5-10-11-12-13-1 with learning function (LF) of trainlm, activation function (AF) of logsig and purelin, and learning rate (LR) of 0.5. This architecture has a good accuracy with MSE of 0.0643. The results showed that this model can predict CPO production in 2019.

Bio-inspired neural network with application to license plate recognition: hysteretic ELM approach

2016 ◽  
Vol 36 (2) ◽  
pp. 172-178 ◽  
Author(s):  
Liang Chen ◽  
Leitao Cui ◽  
Rong Huang ◽  
Zhengyun Ren
Keyword(s):  
Neural Network ◽  
Activation Function ◽  
Neural System ◽  
Learning System ◽  
License Plate ◽  
Sigmoid Function ◽  
License Plate Recognition ◽  
Content Type ◽  
Hidden Layer ◽  
Automatic License Plate Recognition

Purpose This paper aims to present a bio-inspired neural network for improvement of information processing capability of the existing artificial neural networks. Design/methodology/approach In the network, the authors introduce a property often found in biological neural system – hysteresis – as the neuron activation function and a bionic algorithm – extreme learning machine (ELM) – as the learning scheme. The authors give the gradient descent procedure to optimize parameters of the hysteretic function and develop an algorithm to online select ELM parameters, including number of the hidden-layer nodes and hidden-layer parameters. The algorithm combines the idea of the cross validation and random assignment in original ELM. Finally, the authors demonstrate the advantages of the hysteretic ELM neural network by applying it to automatic license plate recognition. Findings Experiments on automatic license plate recognition show that the bio-inspired learning system has better classification accuracy and generalization capability with consideration to efficiency. Originality/value Comparing with the conventional sigmoid function, hysteresis as the activation function enables has two advantages: the neuron’s output not only depends on its input but also on derivative information, which provides the neuron with memory; the hysteretic function can switch between the two segments, thus avoiding the neuron falling into local minima and having a quicker learning rate. The improved ELM algorithm in some extent makes up for declining performance because of original ELM’s complete randomness with the cost of a litter slower than before.

Sign in / Sign up

Export Citation Format

Share Document