Linear Radial Basis Function Research Articles

Comparisons of Speech Parameterisation Techniques for Classification of Intellectual Disability Using Machine Learning

Classification of intellectually disabled children through manual assessment of speech at an early age is inconsistent, subjective, time-consuming and prone to error. This study attempts to classify the children with intellectual disabilities using two speech feature extraction techniques: Linear Predictive Coding (LPC) based cepstral parameters, and Mel-frequency cepstral coefficients (MFCC). Four different classification models: k-nearest neighbour (k-NN), support vector machine (SVM), linear discriminant analysis (LDA) and radial basis function neural network (RBFNN) are employed for classification purposes. 48 speech samples of each group are taken for analysis, from subjects with a similar age and socio-economic background. The effect of the different frame length with the number of filterbanks in the MFCC and different frame length with the order in the LPC is also examined for better accuracy. The experimental outcomes show that the projected technique can be used to help speech pathologists in estimating intellectual disability at early ages.

Improving the Accuracy of the Video Popularity Prediction Models through User Grouping and Video Popularity Classification

This article proposes a novel approach for enhancing the video popularity prediction models. Using the proposed approach, we enhance three popularity prediction techniques that outperform the accuracy of the prior state-of-the-art solutions. The major components of the proposed approach are two novel mechanisms for ” user grouping ” and ” content classification .” The user grouping method is an unsupervised clustering approach that divides the users into an adequate number of user groups with similar interests. The content classification approach identifies the classes of videos with similar popularity growth trends. To predict the popularity of the newly-released videos, our proposed popularity prediction model trains its parameters in each user group and its associated video popularity classes. Evaluations are performed through a 5-fold cross validation and on a dataset containing one month video request records of 26,706 users of BBC iPlayer. Using the proposed grouping technique, user groups of similar interest and up to two video popularity classes for each user group were detected. Our analysis shows that the accuracy of the proposed solution outperforms the state-of-the-art, including Szabo-Huberman (SH), Multivariate Linear (ML), and Multivariate linear Radial Basis Functions (MRBF) models by an average of 45%, 33%, and 24%, respectively. Finally, we discuss how various systems in the network and service management domain such as cache deployment, advertising, and video broadcasting technologies benefit from our findings to illustrate the implications.

Optimal estimation of physical properties of the products of an atmospheric distillation column using support vector regression

Atmospheric distillation column is one of the most important units in an oil refinery where crude oil is fractioned into its more valuable constituents. Almost all of the state-of-the art online equipment has a time lag to complete the physical property analysis in real time due to complexity of the analyses. Therefore, estimation of the physical properties from online plant data with a soft sensor has significant benefits. In this paper, we estimate the physical properties of the hydrocarbon products of an atmospheric distillation column by support vector regression using Linear, Polynomial and Gaussian Radial Basis Function kernels and SVR parameters are optimized by using a variety of algorithms including genetic algorithm, grid search and non-linear programming. The optimization-based data analytics approach is shown to produce superior results compared to linear regression, the mean testing error of estimation is improved by 5% with SVR 4.01 °C to 3.8 °C.

APSO based LLRBFNN Model and Enhanced Fuzzy C Means algorithm for Brain Tumor Detection and Classification from Magnetic Resonance Image

This paper presents a novel APSO (Accelerated Particle Swarm Optimization) Predicated LLRBFNN (Local Linear Radial Basis Function Neural Network) model for automatic encephalon tumor detection and classification. The enhanced fuzzy c means algorithm (EnFCM) has been proposed for image segmentation and the GLCM (Gray Level Co-occurrence Matrix) technique for feature extraction from MR (Magnetic Resonance) images. This research work aims to utilize the hybrid models and algorithms for relegation and segmentation of encephalon tumors from the MR images. The extracted features have been alimented as input to the proposed APSO predicated LLRBFNN model for relegation of benign and malignant tumors. In this research work the proposed LLRBFNN model weights are optimized by utilizing APSO training which will provide unique solution to mitigation the hectic task of radiologist from manual detection of encephalon tumors from MR Images. Additionally the centers of the LLRBFNN model are culled by the Enhanced Fuzzy C Means algorithm and updated by the APSO algorithm. The results of proposed PSO predicated LLRBFNN model has been compared with PSO-LLRBFNN model, APSO-RBFNN and PSO-RBFNN model and the comparison results are presented. The experimental results obtained from the proposed model shows better relegation results as compared to the subsisting models proposed anteriorly.

Effect of load, sliding distance and sliding velocity on the wear properties of aluminum alloy AA5052

Aluminum alloys are widely used in engineering applications. In motion established contact applications, wear is an inevitable phenomenon. In this study, the wear mechanism of AA5052was explored using pin-on-disc tribometer. The wear test parameters namely load (kg), sliding distance (m), and velocity (m/s) were varied according to central composite design. The wear tracks of the worn specimens were observed using high-resolution scanning electron microscope and the elemental composition was analysed using energy dispersive X-ray spectroscopy. A hybrid model integrating the linear function and radial basis function was developed to explore the effect of load, sliding distance, and sliding velocity on the wear rate of the AA5052 alloy. The results indicate that increase in axial load and sliding distance decreases the wear rate of the AA5052 alloy.

An improved hybrid-ORBIT algorithm based on point sorting and MLE technique

Optimization using radial basis functions as an interpolation tool in trust region (ORBIT) is a derivative-free framework based on fully linear radial basis function (RBF) models. In this paper, an improved version of ORBIT algorithm based on two novel ideas is proposed. The accuracy and stability of RBFs depend on a so-called shape parameter, so it is more appropriate to determine the shape parameter according to the optimization problem. While ORBIT in all problems uses a fixed value as a shape parameter, our new version, Hybrid-ORBIT, uses a statistical technique to select an appropriate shape parameter. In addition, ORBIT uses some stored points to build a fully linear RBF model without considering their function values, while in the Hybrid-ORBIT algorithm the stored points are sorted based on their function values and the RBF model is built using the points with lower function values, and the best point in the sense of function value is defined as the trust-region center. Numerical results indicate the efficiency of the improved version compared with the original version.

Effect of composition and aging time on hardness and wear behavior of Cu-Ni-Sn spinodal alloy

Copper alloyed with various compositions of nickel and tin were cast into molds under argon atmosphere. The cast rods were homogenized, solution heat treated, followed by aging for different time duration. The specimens were characterized for microstructure and tested for microhardness and wear rate. A hybrid model with a linear function and radial basis function was developed to analyze the influence of nickel, tin, and aging time on the microhardness and tribological behavior of copper-nickel-sin alloy system. The results indicate that increase in the composition of nickel and tin increases the microhardness and decreases the wear rate of the alloy. The increase in the concentration of nickel and tin decreases the peak aging time of the alloy system.

Performance of Support Vector Machine Kernels (SVM-K) on Breast Cancer (BC) Dataset

Breast cancer (BC) most diagnosed invasive disorder and important cause of casualty for women worldwide. Indian contest BC most commonly spread disease among females. This problem is more alarming to economically developing country like India. Government of India made a lot of effort to make aware the women of the country, but despite of availability of diagnostic tool, prediction of disease in real situation is still a puzzle for researchers. Timely detection and categorization of BC using the evolving techniques like Machine Learning (ML) can show a significant role in BC identification and this could be a preventive policy which effectively reduces the risk of BC patients. Although there are four Kernels in ML, are widely in use but their performance varies with the kind of data available. In this study we, apply four different Kernels such as Linear Kernel (LK), Polynomial Kernel (PK), Sigmoid Kernel (SK) and Radial Basis Function Kernel (RBFK) on BC dataset. We estimated the performance of Support Vector Machine Kernels (SVM-K) on BC dataset .The basic idea is to check the exactness of SVM-K to classify WBCD in terms of effectiveness with respect to accuracy, runtime, specificity and precision. The investigations outcome displays that RBFK provides greater accuracy with minimal errors

MASCA–PSO based LLRBFNN model and improved fast and robust FCM algorithm for detection and classification of brain tumor from MR image

A novel modified adaptive sine cosine optimization algorithm (MASCA) integrated with particle swarm optimization (PSO) based local linear radial basis function neural network (LLRBFNN) model has been proposed for automatic brain tumor detection and classification. In the process of segmentation, the fuzzy C means algorithm based techniques drastically fails to remove noise from the magnetic resonance images. So, for reduction of noise and smoothening of brain tumor magnetic resonance image an improved fast and robust fuzzy c means algorithm segmentation algorithm has been proposed in this research work. The gray level co-occurrence matrix technique has been employed to extract features from brain tumor magnetic resonance images and the extracted features are fed as input to the proposed modified ASCA–PSO based LLRBFNN model for classification of benign and malignant tumors. In this research work the LLRBFNN model’s weights are optimized by using proposed MASCA–PSO algorithm which provides a unique solution to get rid of the hectic task of radiologist from manual detection. The classification accuracy results obtained from sine cosine optimization algorithm, PSO and adaptive sine cosine optimization algorithm integrated with particle swarm optimization based LLRBFNN models are compared with the proposed MASCA–PSO based LLRBFNN model. It is observed that the result obtained from the proposed model shows better classification accuracy results as compared to the other LLRBFNN based models.

Smooth Support Vector Machine (SSVM) for classification of Human Development Index

Human Development Index (HDI) is measuring achievements of human development based on basic components of quality of life. Human development index is low score if HDI is less than 60, moderate HDI between 60 to less than 70, high HDI between 70 to less than 80, and equal to 80 and more than 80 belong to high HDI. Smooth Support Vector Machine (SSVM) is a classification technique that is new. The algorithm used is Newton Armijo with linear kernel, polynomial kernel, and Radial Basis Function (RBF) kernel. The result of classification of human development index with SSVM method with linear kernel shows the prediction accuracy 84.77%, polynomial kernel 61.65%, and RBF kernel 100%. Radial Base Function Kernel (RBF) is the most accurate kernel in predicting human development index.

Multiple Modeling Techniques for Assessing Sesame Oil Extraction under Various Operating Conditions and Solvents.

This paper compares four different modeling techniques: Response Surface Method (RSM), Linear Radial Basis Functions (LRBF), Quadratic Radial Basis Functions (QRBF), and Artificial Neural Network (ANN). The models were tested by monitoring their performance in predicting the optimum operating conditions for Sesame seed oil extraction yields. Experimental data using three different solvents—hexane, chloroform, and acetone—with varying ratios of solvents to seeds, all under different temperatures, rotational speeds, and mixing times, were modeled by the three proposed techniques. Efficiency for model predictions was examined by monitoring error value performance indicators (R2, R2adj, and RMSE). Results showed that the applied modeling techniques gave good agreements with experimental data regardless of the efficiency of the solvents in oil extraction. On the other hand, the ANN model consistently performed more accurate predictions with all tested solvents under all different operating conditions. This consistency is demonstrated by the higher values of R2 and R2adj ratio equals to one and the very low value of error of RMSE (2.23 × 10−3 to 3.70 × 10−7), thus concluding that ANN possesses a universal ability to approximate nonlinear systems in comparison to other models.

LTV-RBF Approach for Yaw Stability Control of Distributed Drive Electric Vehicles

This paper proposed a torque distribution strategy based on linear time-varying radial basis function (LTVRBF) neural networks for yaw stability control of electric vehicles equipped with in-wheel motors. The desired yaw rate is calculated by the two-degree-of-freedom (2- DOF) bicycle dynamic model. The influence of the timevarying steering angle is considered for reducing the yaw rate error. To solve this problem, the constant connection weight of conventional RBF networks is converted into a time-varying variable which is used to track the reference trajectory and optimize the torque distribution. The torques of in-wheel motors are restricted in high efficiency range to improve the system energy efficiency. Simulation results of the proposed torque distribution strategy based on dSPACE simulator show that LTV-RBF networks can effectively track the reference yaw rate and stabilize the system.

An application of local linear radial basis function neural network for flood prediction

Heavy seasonal rain makes waterway flood and is one of the preeminent reason behind flooding. Flooding causes various perils with outcomes including danger to human life, harm to building, streets, misfortune to horticultural fields and bringing about human uprooting. Thus, prediction of flood is of prime importance so as to reduce exposure of people and destruction of property. This paper focuses on applying different neural networks approach, i.e. Multilayer Perceptron, Radial Basis functional neural network, Local Linear Radial Basis Functional Neural Network and Artificial Neural Network with Whale Optimization to predict flood in terms of rainfall, gauge, area, velocity, pressure, average temperature, average wind speed that are setup through field and lab investigation from the contextual analysis of river “Daya” and “Bhargavi”. It has always been a troublesome undertaking to predict flood as many factors have influence on it although with this neural network models the prediction accuracy can be optimized using back propagation method which is a widely applied over traditional learning method for neural system because of its preeminent learning ability. The flood prediction system is built with the four models and a comparison is made which provides us the answer to which model is effective for the prediction.

Energy Time Series Forecasting Based on Empirical Mode Decomposition and FRBF-AR Model

Energy time series forecasting is challenging due to its uncertainty of natural factor. Energy time series has nonlinear and non-stationary characteristics and not normally distributed, which makes it difficult to be forecast by statistical or computational intelligent methods. A set of locally linear fuzzy radial basis function (FRBF) to approximate the functional coefficients of the state-dependent autoregressive (SD-AR) model and the proposed model is called FRBF-AR model, which combines the advantage of FRBF in function approximation and the merit of SD-AR model in nonlinear dynamics description. Empirical mode decomposition (EMD) is a powerful tool to decompose a complex time series into a number of intrinsic mode functions (IMFs) and one residual series. Then, these sub-series are predicted by the FRBF-AR model. Finally, the prediction results of the IMFs and residual series are added to formulate an ensemble forecast for the original energy time series. Furthermore, energy time series data sets from the generated electricity and load demand time series are used to test the effectiveness of the proposed hybrid modeling method. The simulation results demonstrate the effectiveness of the proposed hybrid modeling method compared with other forecasting methods.

Prediction of the Toxicity of Binary Mixtures by QSAR Approach Using the Hypothetical Descriptors.

Organic compounds are often exposed to the environment, and have an adverse effect on the environment and human health in the form of mixtures, rather than as single chemicals. In this paper, we try to establish reliable and developed classical quantitative structure–activity relationship (QSAR) models to evaluate the toxicity of 99 binary mixtures. The derived QSAR models were built by forward stepwise multiple linear regression (MLR) and nonlinear radial basis function neural networks (RBFNNs) using the hypothetical descriptors, respectively. The statistical parameters of the MLR model provided were N (number of compounds in training set) = 79, R2 (the correlation coefficient between the predicted and observed activities)= 0.869, LOOq2 (leave-one-out correlation coefficient) = 0.864, F (Fisher’s test) = 165.494, and RMS (root mean square) = 0.599 for the training set, and Next (number of compounds in external test set) = 20, R2 = 0.853, (leave-one-out correlation coefficient for test set)= 0.825, F = 30.861, and RMS = 0.691 for the external test set. The RBFNN model gave the statistical results, namely N = 79, R2 = 0.925, LOOq2 = 0.924, F = 950.686, RMS = 0.447 for the training set, and Next = 20, R2 = 0.896, = 0.890, F = 155.424, RMS = 0.547 for the external test set. Both of the MLR and RBFNN models were evaluated by some statistical parameters and methods. The results confirm that the built models are acceptable, and can be used to predict the toxicity of the binary mixtures.

Estimation of rock uniaxial compressive strength for an Iranian carbonate oil reservoir: Modeling vs. artificial neural network application

Estimation of rock mechanic parameters is an important issue in reservoir management. Uniaxial compressive strength (UCS) and elastic modulus are the most important factors in determining the rock mechanic parameters in petroleum engineering studies. Accessibility to the parameters in fields such as designing fracture, analyzing of wellbore stability and drilling programming are very useful. The most accurate method to assign the aforementioned parameters is measuring these parameters in a laboratory. Laboratory determination of these parameters is problematic work due to technology issues, lack of laboratory equipment and coring problems in oil and gas wells, so indirect estimation of these parameters is required. Using well log data is the cheapest and most available approach in order to indirectly estimate these parameters. In this investigation, different models including multiple linear regression (MLR) and artificial neural network (ANN) (i.e., multi linear perceptron (MLP) and radial basis function (RBF)) were utilized for prediction of UCS via the three parameters of porosity, density and water saturation. These data were obtained from analysis of sonic, neutron, gamma ray and electric logs. The best results were obtained from a 3-15-1 MLP network which included one hidden layer and 15 neurons from the hidden layer using the trial and error method, and a 3-17-1 RBF which included 17 hidden neurons and a spread ó of 1.6. The core data from one of the carbonate Iranian oil fields (Asmari reservoir) were utilized for training, validation and testing of the networks, and correlation coefficients of 0.68, 0.90 and 0.83 were obtained for MLR, MLP and RBF, respectively.

MHD convection ow in a constricted channel

Abstract We consider the steady, laminar, convection ow in a long channel of 2D rectangular constricted cross-section under the inuence of an applied magnetic field. The Navier-Stokes equations including Lorentz and buoyancy forces are coupled with the temperature equation and are solved by using linear radial basis function (RBF) approximations in terms of the velocity, pressure and the temperature of the fluid. RBFs are used in the approximation of the particular solution which becomes also the approximate solution of the problem. Results are obtained for several values of Grashof number (Gr), Hartmann number (M) and the constriction ratios (CR) to see the effects on the ow and isotherms for fixed values of Reynolds number and Prandtl number. As M increases, the ow is flattened. An increase in Gr increases the magnitude of the ow in the channel. Isolines undergo an inversion at the center of the channel indicating convection dominance due to the strong buoyancy force, but this inversion is retarded with the increase in the strength of the applied magnetic field. When both Hartmann number and constriction ratio are increased, ow is divided into more loops symmetrically with respect to the axes.

Estimation of the Toxicity of Different Substituted Aromatic Compounds to the Aquatic Ciliate Tetrahymena pyriformis by QSAR Approach.

Nowadays, quantitative structure–activity relationship (QSAR) methods have been widely performed to predict the toxicity of compounds to organisms due to their simplicity, ease of implementation, and low hazards. In this study, to estimate the toxicities of substituted aromatic compounds to Tetrahymena pyriformis, the QSAR models were established by the multiple linear regression (MLR) and radial basis function neural network (RBFNN). Unlike other QSAR studies, according to the difference of functional groups (−NO2, −X), the whole dataset was divided into three groups and further modeled separately. The statistical characteristics for the models are obtained as the following: MLR: n = 36, R2 = 0.829, RMS (root mean square) = 0.192, RBFNN: n = 36, R2 = 0.843, RMS = 0.167 for Group 1; MLR: n = 60, R2 = 0.803, RMS = 0.222, RBFNN: n = 60, R2 = 0.821, RMS = 0.193 for Group 2; MLR: n = 31 R2 = 0.852, RMS = 0.192; RBFNN: n = 31, R2 = 0.885, RMS = 0.163 for Group 3, respectively. The results were within the acceptable range, and the models were found to be statistically robust with high external predictivity. Moreover, the models also gave some insight on those characteristics of the structures that most affect the toxicity.

Cycle reservoir with regular jumps for forecasting ozone concentrations: two real cases from the east of Croatia

Satisfying the national air quality standards represents a challenge nowadays for developing countries. Air pollution in industrial cities is one of the foremost problems that affect human health and might cause loss of human life. One of the main attributes that can cause a significant impact on people’s health is the ground-level ozone pollution. Ozone can raise the ratio of asthma attacks, permanent damage to lungs, and maybe death. Forecasting its concentration levels is essential for planning well-designed environment protection strategies. In this paper, a state-space reservoir model called cycle reservoir with jumps (CRJ) is used to predict the level of ozone concentrations in the east of Croatia utilizing some meteorological parameters including the temperature, relative humidity, wind speed, wind direction, and the pollutants PM10. CRJ is a particular type of recurrent neural networks with powerful performance when applied for complex temporal problems. Two cases from the east of Croatia are investigated in this work: the Kopacki Rit area and the Osijek city. The proposed CRJ model shows superiority of CRJ model in forecasting ozone concentrations compared to linear regression, multilayer perceptron (MLP) and radial basis function (RBF) network.

The identification of high potential archers based on relative psychological coping skills variables: A Support Vector Machine approach

Support Vector Machine (SVM) has been revealed to be a powerful learning algorithm for classification and prediction. However, the use of SVM for prediction and classification in sport is at its inception. The present study classified and predicted high and low potential archers from a collection of psychological coping skills variables trained on different SVMs. 50 youth archers with the average age and standard deviation of (17.0 ±.056) gathered from various archery programmes completed a one end shooting score test. Psychological coping skills inventory which evaluates the archers level of related coping skills were filled out by the archers prior to their shooting tests. k-means cluster analysis was applied to cluster the archers based on their scores on variables assessed. SVM models, i.e. linear and fine radial basis function (RBF) kernel functions, were trained on the psychological variables. The k-means clustered the archers into high psychologically prepared archers (HPPA) and low psychologically prepared archers (LPPA), respectively. It was demonstrated that the linear SVM exhibited good accuracy and precision throughout the exercise with an accuracy of 92% and considerably fewer error rate for the prediction of the HPPA and the LPPA as compared to the fine RBF SVM. The findings of this investigation can be valuable to coaches and sports managers to recognise high potential athletes from the selected psychological coping skills variables examined which would consequently save time and energy during talent identification and development programme.