Application Of Probabilistic Neural Network Research Articles

Probabilistic neural networks for incremental learning over time-varying streaming data with application to air pollution monitoring

This paper proposes a novel algorithm for incremental learning over streaming data in a non-stationary environment. The idea refers to the applicability of Probabilistic Neural Networks (PNNs), commonly used as a fast and robust method for solving classification problems in a stationary scenario. It is well known that PNNs have solid mathematical foundations but they fail in the non-stationary scenario and are not able to deal with concept drift. In this paper, the method based on the orthogonal series expansions of unknown probability densities is proposed. This approach significantly extends the applicability of the PNNs. The nonparametric procedures called Incremental Probabilistic Neural Networks (IPNNs) for tracking drifting probability densities or conditional class densities, in the case of classification, are presented. We prove their convergence as the stream data size is growing to infinity. More precisely, we show convergence in probability and with probability one as the sample size tends to infinity. Compared to the existing literature, almost exclusively based on various heuristics, the proposed method is mathematically justified and has solid theoretical foundations. Simulations performed on synthetic and air pollution data confirm the effectiveness of the algorithm.

Application of probabilistic neural network for speech emotion recognition

Application of probabilistic neural network for speech emotion recognition

Reducing the number of centers in a probabilistic neural network via applying the first neighbor means clustering algorithm

Probabilistic neural network is a variant of feedforward neural network models and has been successfully applied for various pattern classification purposes. Unlike other feedforward neural network models, probabilistic neural network, a type of radial basis function network models, has essentially only two types of the network-parameters to choose in advance, i.e. the locations of the centers and a single value of radius; a central issue relevant to the application of probabilistic neural network is therefore to determine an appropriate number of the centers accommodated within the network. In the original probabilistic neural network framework, all the training data are allocated to the respective centroid vectors, and thus the network size generally tends to be large, resulting in demanding computational resource. To alleviate this problem, clustering algorithms are commonly employed to shrink the size of the training data. In this work, reduction in the number of centers in a probabilistic neural network is addressed, via the utility of first neighbor means clustering algorithm that is non-iterative and requires only a single algorithmic hyper-parameter; such a choice is desirable in practice. Simulation results using seven publicly available databases for pattern classification tasks show that the first neighbor means clustering algorithm can yield a relatively compact-sized network within short computation time, while exhibiting a reasonably high classification performance, in comparison with communication with local agents, k-means, orthogonal least squares, resource allocating network, and resource vector machine algorithms.

Application of Probabilistic Neural Networks Using High-Frequency Components’ Differential Current for Transformer Protection Schemes to Discriminate between External Faults and Internal Winding Faults in Power Transformers

Internal and external faults in a power transformer are discriminated in this paper using an algorithm based on a combination of a discrete wavelet transform (DWT) and a probabilistic neural network (PNN). DWT decomposes high-frequency fault components using the maximum coefficients of a ¼ cycle DWT as input patterns for the training process in a decision algorithm. A division algorithm between a zero sequence of post-fault differential current waveforms and the differential current coefficient in the ¼ cycle DWT is used to detect the maximum ratio and faults. The simulation system uses various study cases based on Thailand’s electricity transmission and distribution systems. The simulation results demonstrated that the PNN and BPNN are effectively implemented and perform fault detection with satisfactory accuracy. However, the PNN method is most suitable for detecting internal and external faults, and the maximum coefficient algorithm is the most effective in detecting the fault. This study will be useful in differential protection for power transformers.

Application of probabilistic neural network with transmission and distribution protection schemes for classification of fault types on radial, loop, and underground structures

This paper proposes the development of transmission and distribution protection schemes to classify faults along transmission and distribution systems. The systems under consideration are composed of a 500-kV radial (two-bus single circuit), loop (three-bus double circuit) structure transmission systems, and a 115-kV radial structure underground distribution system. This complex system shows the advantage of the proposed method. A decision algorithm based on a discrete wavelet transform (DWT) and probabilistic neural network is investigated for inclusion in a transmission and distribution protection system. Fault signals in each case are extracted to several scales on the DWT to decompose high-frequency components from fault signals using the mother wavelet daubechies4. The maximum coefficients of a DWT at 1/4 cycle that can detect faults are used as input patterns for the training process in a decision algorithm. In addition, coefficients from a DWT technique and a back-propagation neural network algorithm are also compared with the proposed algorithm in this paper. Moreover, the real signal from an experimental set-up was investigated. Based on the accurate results for the simulation signal and real signal, it can be concluded that the proposed algorithm is adequate for other power systems with different line models and that it can be applied to actual systems even if the accuracy is slightly reduced by the effect of noise in an actual system. Thus, the overall results show that the proposed algorithm can detect a faulty bus and classify types of fault with satisfactory results.

SDSS-DR12 bulk stellar spectral classification: Artificial neural networks approach

This paper explores the application of Probabilistic Neural Network (PNN), Support Vector Machine (SVM) and Kmeans clustering as tools for automated classification of massive stellar spectra.

Improved probabilistic neural network PNN and its application to defect recognition in rock bolts

The probabilistic neural network (PNN) model is developed based on Bayesian criteria of multivariate pattern classification. Because PNN has the advantages of low training complexity, high stability, quick convergence, and simple construction, it can be well applied in the area of defect recognition. The application of PNN has a bottleneck problem; that is, the selection of smoothing parameters will seriously affect PNN recognition accuracy. In this paper, we propose an improved PNN model that employs a differential evolution algorithm to optimize the smoothing parameters. Furthermore, a defect recognition approach is proposed based on the improved PNN, which is successfully applied to an anchoring testing field. The proposed approach includes two key steps. The first step involves extracting the energy eigenvector of the defect signal with normalization based on a wavelet packet. The second step involves recognizing defects based on the improved PNN. Simulation results show that our improved PNN model is superior to the traditional PNN model. Thus, the improved PNN can provide useful references for recognizing the defect type of rock bolts in engineering.

Automation of Segmentation of MRI Images in the Presence of Intensity Inhomogeneity by the Application of Probabilistic Neural Network

This paper presents a different approach by employing probabilistic neural network (PNN) with level set method for segmentation of intensity inhomogeneous magnetic resonance images (MRI). The input image is classified by PNN, and initialise the image with initial zero level set contours. A cascade approach is used in which after processing image by PNN, the output of PNN is fed into Level Set Method (LSM) Algorithm. The LSM is usedto perform the segmentation and to estimate bias field.

Application of Probabilistic Neural Networks to microhabitat suitability modelling for adult brown trout (Salmo trutta L.) in Iberian rivers

Probabilistic Neural Networks (PNN) have been tested for the first time in microhabitat suitability modelling for adult brown trout (Salmo trutta L.). The impact of data prevalence on PNN was studied. The PNN were evaluated in an independent river and the applicability of PNN to assess the environmental flow was analysed. Prevalence did not affect significantly the results. However PNN presented some limitations regarding the output range. Our results agreed previous studies because trout preferred deep microhabitats with medium-to-coarse substrate whereas velocity showed a wider suitable range. The 0.5 prevalence PNN showed similar classificatory capability than the 0.06 prevalence counterpart and the outputs covered the whole feasible range (from 0 to 1), but the 0.06 prevalence PNN showed higher generalisation because it performed better in the evaluation and it allowed a better modulation of the environmental flow. PNN has demonstrated to be a tool to be into consideration.

Application of Probabilistic Neural Network in Pattern Classification

The network model of probabilistic neural network and its method of pattern classification and discrimination are first introduced in this paper. Then probabilistic neural network and three usually used back propagation neural networks are established through MATLAB7.0. The pattern classification of dots on a two-dimensional plane is taken as an example. Probabilistic neural network and improved back propagation neural networks are used to classify these dots respectively. Their classification results are compared with each other. The simulation results show that compared with back propagation neural networks, probabilistic neural network has simpler learning rules, faster training speed and it needs fewer training samples; the pattern classification method based on probabilistic neural network is very effective, and it is superior to the one based on back propagation neural networks in classifying speed, accuracy as well as generalization ability.

Application of Probabilistic Neural Networks in Fault Diagnosis of Three-Phase Induction Motors

In fault diagnosis of three-phase induction motors, traditional methods usually fail because of the complex system of three-phase induction motors. Short circuit is a very common stator fault in all the faults of three-phase induction motors. Probabilistic neural network is a kind of artificial neural network which is widely used due to its fast training and simple structure. In this paper, the diagnosis method based on probabilistic neural network is proposed to deal with stator short circuits. First, the principle and structure of probabilistic neural network is studied in this paper. Second, the method of fault setting and fault feature extraction of three-phase induction motors is proposed on the basis of the fault diagnosis of stator short circuits. Then the establishment of the diagnosis model based on probabilistic neural network is illustrated with details. At last, training and simulation tests are done for the model. And simulation results show that this method is very practical with its high accuracy and fast speed.

Application of probabilistic neural network in bacterial identification by biochemical profiles

An algorithm of Probabilistic Neural Network (PNN) for bacterial identification based on the probability matrix of API 20E system as a case study is reported. The PNN shows the correct identification of all the taxa belonging to the training and test sets and possesses merits over the conventional methods.

The Application of Probabilistic Neural Network in Speech Recognition Based on Partition Clustering

A probabilistic neural network (PNN) speech recognition model based on the partition clustering algorithm is proposed in this paper. The most important advantage of PNN is that training is easy and instantaneous. Therefore, PNN is capable of dealing with real time speech recognition. Besides, in order to increase the performance of PNN, the selection of data set is one of the most important issues. In this paper, using the partition clustering algorithm to select data is proposed. The proposed model is tested on two data sets from the field of spoken Arabic numbers, with promising results. The performance of the proposed model is compared to single back propagation neural network and integrated back propagation neural network. The final comparison result shows that the proposed model performs better than the other two neural networks, and has an accuracy rate of 92.41%.

Comparison between Different Classification Methods with Application to Skin Cancer

recent years, skin cancer is the most common form of human cancer. It is estimated that over 1 million new cases occur annually. In order to detect skin cancer various methods have been proposed in the past decades. This paper focuses on the development of a skin cancer screening system that can be used in a general practice by non-experts to classify normal from abnormal cases. The development process consists of Feature Detection and Classification Technique. The features are extracted by decomposing images into different frequency sub- bands using wavelet transform. The output of Discrete Wavelet Transform becomes input to the Classification System which classify whether the input image is cancerous or noncancerous. The classification system is based on the application of Probabilistic Neural Network and Clustering Classifier. The Accuracy of the proposed system is calculated using different classification techniques on image database of 80 samples (40 cancerous and 40 non cancerous images).

Discrete harmony search based expert model for epileptic seizure detection in electroencephalography

Seizure detection and classification using signal processing methods has been an important issue of research for the last two decades. In the present study, a novel scheme was presented to detect epileptic seizure activity with very fast and highest accuracy from background electro encephalogram (EEG) data recorded from epileptic and normal subjects. The proposed scheme is based on discrete wavelet packet transform (DWT) with energy, entropy, standard deviation, mean, kurtosis, skewness and entropy estimation at each node of the decomposition tree followed by application of probabilistic neural network (PNN). Normal as well as epileptic EEG epochs were decomposed into approximation and details coefficients till sixth-level using DWT packet. Discrete harmony search with modified differential operator was used to select the optimal features out of all above mentioned statistical and non-statistical parameters. In order to demonstrate the efficacy of the proposed algorithm for classification purpose using PNN, we have implemented 10-fold cross validation. Clinical EEG data recorded from normal as well as epileptic subjects are used to test the performance of this new scheme. It is found that the detection rate is 100% accurate with same level of sensitivity and specificity.

Application of Probabilistic Neural Network in Bonding Quality Ultrasonic Detection of Composite Material

Application of Probabilistic Neural Network in Bonding Quality Ultrasonic Detection of Composite Material

Probabilistic neural network approach to the classification of demonstrative pronouns for indirect anaphora in Hindi

In this paper, we propose the application of probabilistic neural networks (PNNs) to the classification scheme of demonstrative pronouns for indirect anaphora in Hindi corpus. The Demonstrative Pronouns in Hindi, “ yeh”(this/it), “ veh”(that/those), “ iss”(this/it), and “ uss”(that/those) can be personal or demonstrative. The differentiation can be ascertained from only the situation or the context. The case marking of pronouns further add the constraints on linguistic patterns. We propose to cast such an anaphora as a semantic inference process, which encompasses several salient linguistic characteristic features such as grammatical role, proximity, syntactic category and semantic cues. Our focus of study is demonstrative pronouns without noun phrase antecedent in Hindi written corpus. We analyzed 313 news items having 3890 sentences, 3101 pronouns, of which 608 instances covered those demonstrative pronouns, which had 183 instances with non-NP-antecedents. The effectiveness of the approach is demonstrated through set of simulations and evaluations.

Abstracts

One of the major problems facing Blast Furnaces is the occurrence of cracks in taphole mud, as the underlying causes are not easily identifiable. The absence of this knowledge makes it difficult the use of conventional techniques for predictability and mitigation. This paper will address the application of Probabilistic Neural Network using the Matlab software as a means to detect and control such cracks. The most relevant BF operational variables were picked through the statistic tool Principal Component Analysis - PCA. Based upon the selection of these variables a probabilistic neural network was built. A set of BF operational data, consisting of 30 controlling variables, was divided into 2 groups, one of which for network training, and the other one to validate the neural network. The neural network got 98% of the cases right. The results show the effectiveness of this tool for crack prediction in relation to clay intrinsic properties and as a result of the fluctuation in operational variables.

Expert model for detection of epileptic activity in EEG signature

Seizure detection and classification using signal processing methods has been an important issue of research for the last two decades. In the present study, a novel scheme was presented to detect epileptic seizure activity with very fast and high accuracy from background electro encephalogram (EEG) data recorded from epileptic and normal subjects. The proposed scheme is based on discrete wavelet transform (DWT) and energy estimation at each node of the decomposition tree followed by application of probabilistic neural network (PNN) for classification. Normal as well as epileptic EEG epochs were decomposed into approximation and details coefficients till the sixth-level using DWT. Approximate energy (EDA) values of the wavelet coefficients at all nodes of the down sampled tree were used as a feature vector to characterize the predictability of the epileptic activity within the records of EEG data. In order to demonstrate the classification accuracy of the proposed probabilistic neural network, tenfold cross-validation was implemented in the expert model. Clinical EEG data recorded from normal as well as epileptic subjects were used to test the performance of this new scheme. It was found that with the proposed scheme, the detection is 99.33% accurate with sensitivity and specificity as 99.6% and 99%, respectively. The proposed model can be widely used in developing countries where there is an acute shortage of trained neurologist.

Probabilistic neural networks applied to mineral potential mapping for platinum group elements in the Serra Leste region, Carajás Mineral Province, Brazil

This work presents an application of probabilistic neural networks to map the potential for platinum group elements (PGE) mineralization sites in the northeast portion of the Carajás Mineral Province (CMP), Brazilian Amazon. Geological and geophysical gamma-spectrometric and magnetic data were used to generate evidential maps to derive input feature vectors. Feature vectors representing known mineralized locations were used as training data. The networks were created based on the training dataset and the evidential maps were classified in terms of probabilities using these networks. We have produced mineral potential models that depict classes with high, moderate and low favorability for Au-PGE mineralization sites and a model with high and low favorability classes for Cr-PGE mineralization sites. The cut-off values for each class were selected as the inflexion points of the curves of favorability against cumulative percentage of the study area. These curves were also used to check for the efficiency of the models by plotting the favorability values at the training sites. Leave-one-out tests were applied to validate the models and the overall accuracy is 87.5%. For Au-PGE mineralization sites, the high favorability areas accounts for 0.57% of the study area and are comprised mainly within meta-pelites and meta-siltites. For Cr-PGE mineralization sites, the high favorability areas are much more restrict and accounts for only 0.17% of the study area, being associated chiefly with mafic and ultramafic rocks. These mineral potential maps can be used as reconnaissance guides for future detailed ground surveys of possible new PGE occurrences, which is of critical importance to shorten exploration time and costs in such densely forested Amazonian terrains.