Search

The paper presents an overview of parameter estimation of three phase induction motor using different techniques. The techniques mainly described in the paper are conventional techniques and soft computing techniques. The soft computing techniques considered in the paper are fuzzy system, artificial neural network (ANN), Neuro-Fuzzy, genetic algorithms (GA) and particle swarm optimization (PSO). [ABSTRACT FROM AUTHOR]

The electroencephalogram (EEG) based Brain-Computer Interfaces (BCIs) has been considerably extended to the multidisciplinary and challenging nature of BCI study. The algorithms of Digital signal processing and pattern recognition certainly constitute necessary components of a BCI system which are applied to the EEG signals to decode mental states those are relevant for BCI operation. The basic BCI concepts such as: brain activity monitoring, BCI operation, and the relevant mental states for BCI are applied to WSN. The major types of relevant mental states for BCI, namely motor imagery (ERD/ERS), steady state visual evoked potentials (SSVEP) are also mapped to MANET's behaviour. The multivariate nature of the EEG combined with the neuroscience familiarity on hemispheric brain specialization is favourably taken into account to derive optimal combinations of the individual signals composing the EEG. BCI provides a non-muscular channel for sending commands to the external world using the electroencephalographic activity or other electrophysiological process of the brain function. A necessary factor in the successful operation of BCI systems is the methods used to process the brain signals. The processed data has been applied on to nodes of MANET where each set of sensor nodes are used for particular application. Based on the output of BCI commands, the MANET is operated and controlled with using only activities of neurons of brain, hence this approach could be more helpful to physically differently abled people or medically injured people. [ABSTRACT FROM AUTHOR]

In the base of the detailed analysis of dam safety monitoring expert system of knowledge, it is introduced that the method of expert system structure of knowledge and the characteristics of knowledge's representation. Maxing the method and ANN to indicate dam safety and the physical structure of the knowledge base is studied. The reasoning strategy based on hybrid knowledge base is also discussed and the technical's problems which forming mixing knowledge base should be paid attention to is analysed. [ABSTRACT FROM AUTHOR]

The difference of reclaimed water qualities between in summer and in winter were proved considerable. Reclaimed water derived in spring or summer could be applied irrigating in some arid region, which have been considered a rational renewable resource of water. Yet in winter, reclaimed water would run off into river or stream in vain. The idea of storing it in winter and being ready for the next year΄s irrigation is proposed, however, the effects on crops are maintained sealed. Experiments processed using raw sewage, primary treated water, and secondary treated water in Gaobeidian Sewage Plant of Beijing to irrigate maize and soybean in spring and summer from 2006 to 2007, based on which the irrigation effect derived from reclaimed water of winter was predicted by ANN. The results validated the strong relations between irrigation effect and qualities of water, which indicated the reclaimed water stored in winter was grossly similar to secondary treated water in summer for irrigation. [ABSTRACT FROM AUTHOR]

In this paper, we discuss the participle type of English sentences in our English to Sanskrit machine translation (EST) system. Our EST system is an integrated model of a rule based machine translation (RBMT) with artificial neural network (ANN) model which translates an English sentence into equivalent Sanskrit sentence. We use feed forward ANN for the selection of Sanskrit word like noun, verb, object, adjective etc from English to Sanskrit user data vector (UDV). Our system uses only morphological markings to identify various part of speech (POS) as well as participle type of sentences. [ABSTRACT FROM AUTHOR]

Paddy is a staple food for more than 93 countries and it will stay of life for future generations. Harvesting is one of the vital operations in crop production and timely harvesting is essential for getting maximum yield. Moisture content and forward speed are the two factors to overcome the post harvest losses and minimise the quantitative losses. In this paper, an artificial neural network is introduced to assess the grain losses in the field condition. The simulation result shows that the ANN method is appropriate and feasible to assess the grain losses. However, model results that, an error of (RMSE) 0.1582 for cutter bar loss, for threshing loss was 0.1299 and for separation loss was 0.1321. Hence, the ANN model help the operator / farmer to decide the time of harvest. It also minimizes the post harvest grain losses by considering crop and machine parameters. [ABSTRACT FROM AUTHOR]

Fingerprint based authentication does not give a foolproof security because fingerprint image can be scanned and can be used later on for the purpose of authentication. In this paper we proposed another level of security by using Session based password in addition to the features of the fingerprint for authentication. The proposed approach shows significant promise and potential for improvement, compared with the other conventional authentication techniques with regard to efficiency of the process. [ABSTRACT FROM AUTHOR]

In this paper, we propose an application tool to help analyze the usage of a mobile phone for a typical user. From the past usage, the tool can analyze the promotion that is suitable for the user which may save the total expense. The application consists of both client and server side. On the server side, the information for each promotion package for a phone operator is stored as well as the usage database for each client. The client side is a user interface for both phone operators and users to enter their information. The analysis engine are based on KNN, ANN, decision tree and Naïve Bayes models. For comparison, it is shown that KNN and decision outperforms the others. [ABSTRACT FROM AUTHOR]

This paper presents a new algorithm for the determination of the Autoregressive Moving Average (ARMA) model order based on Artificial Neural Network (ANN). The basic idea is to apply ANN to a special matrix constructed from the Minimum Eginevalue (MEV) criterion. The MEV criterion is based on a covariance matrix derived from the observed output data only. The input signal is unobservable. The proposed algorithm is based on training the MEV covariance matrix dataset using the back-propagation technique. Our goal is to develop a system based on ANN; hence, the model order can be selected automatically without the need of prior knowledge about the model or any human intervention. Examples are given to illustrate the significant improvement results. [ABSTRACT FROM AUTHOR]

SOAP messages use XML code, which makes them vulnerable to denial of service (DoS) attacks and puts the availability of web services at risk. This article presents an adaptive solution for dealing with DoS attacks in web service environments. The solution proposes a distributed hierarchical multiagent architecture that implements a robust mechanism of classification based on an advanced CBR-BDI agent. The agent incorporates a case-based reasoning engine that integrate a Perceptron Multilayer neural network during the re-use phase to classify incoming SOAP messages and reject those that are considered malicious. A prototype of the architecture was developed and the results obtained are presented in this study. [ABSTRACT FROM AUTHOR]

Selection of characteristic features for a classification task is always crucial to high recognition ratio, regardlessly of the particular processing technique applied. Most methodologies offer some inherent mechanisms of dimension reduction that lead to expression of available data in more succinct way, however, combining elements of distinctively different approaches to data analysis brings interesting conclusions as to the role of particular features and their influence on the power of the resulting classifier. The paper presents research on such fusion of processing techniques, namely employing rough set based analysis of features for ANN classifier within stylometric studies on writing styles. [ABSTRACT FROM AUTHOR]

Aiming to the disadvantages of short-term load forecasting with HHT such as mode mixing and random component, a new short-term load forecasting model based on HHT and ANN is proposed. The first order difference algorithm is adopted to eliminate mode mixing. The random component is forecast with different methods including BP, RBF, SVM, linear and ANN combination. Other components can be forecast with proper method. The simulation results show that the higher accuracy of short-term load forecasting can be obtained. [ABSTRACT FROM AUTHOR]

In this paper, an effective approach is presented to classify epileptic patients from control participants by analyzing their electroencephalogram (EEG) signals. For this aim, first, several time-frequency transforms were applied to five scalp EEG datasets in order to extract discriminant features. Regarding high number of channels and features, kernel-principal-component-analysis (KPCA) was utilized to reduce the feature size in order to decrease the complexity. Then, the projected features were fed to an artificial neural network (ANN) to classify the subjects. The achieved results show that our scheme is capable of classifying normal and epileptic subjects up to 92% and 93% accuracies, respectively. [ABSTRACT FROM AUTHOR]

Empirical Orthogonal Functions (EOF) and Artificial Neural Network (ANN) are performed for investigating and predicting interannual variability in nearshore topographic field around the Yellow River Delta. EOF and ANN are particularly effective at reproducing the observed topographic features around either modern or historical river mouths where the nearshore topography has experienced significantly intense interannual changes during the last thirty decades. In general, the coastal land around the modern river mouths has extended toward the sea and the seafloor has been elevated due to accumulation of a large amount of sediment transported by the modern Yellow River. On the other hand, the coast and the seafloor around the historical river mouths have been eroded quickly by tidal current, waves and storms due to lack of sediment supply. The observed spatial patterns of nearshore topography are well captured qualitatively and quantitatively by dominant eigenvectors of EOF. The EOF principal components indicating temporal variation in dominant eigenvectors are effectively fitted and predicted by using the observed river-related data as the input to ANN. The topographic fields either for the ˵missing″ years or for the ˵future″ years can then be estimated by linear combination of fixed dominant eigenvectors and fitted principal components. As a result, the fitting and predicting errors reach as low as 2.9% and 5.6%, respectively. [ABSTRACT FROM AUTHOR]

Oil well productivity classification and abundance prediction are important for estimating economic benefit of a well. However, it is difficult to predict because well logs are complex and the amount of data collected today has far exceeded our ability to refine and analyze without the use of automated analysis techniques. In response to the problem above mentioned, data mining technology in recent years has shown the ability for discovering information and effectively extracts information from massive observational data sets that can be used to decisions. Especially, classification and prediction methods, are receiving increasing attention from researchers and practitioners in the domain of petroleum exploration and production (E&P) in China. Therefore, data mining is regarded as one of the ten key techniques for challenging problem of oil exploration and development. In this paper, four distinct kinds of classification and prediction methods in data mining, including decision tree (DT), artificial neural network (ANN), support vector machine (SVM) and Bayesian network are used to two real-world case studies. One is hydrocarbon reservoir productivity classification with 21 samples from 16 wells logging data in Karamay Oilfield 8th district reservoir. The results show that SVM and Bayesian are superior in the classification accuracy (95.2%) to DT, ANN and SVM, and can be considered as a prominent classification model. Another is reservoir abundance prediction with 17 mature accumulation systems samples in JiYang depression basin. The results show that SVM is superior in the prediction accuracy (91.92%) to DT, ANN and Bayesian, and can be taken as an excellent prediction model. [ABSTRACT FROM AUTHOR]

A fully automated and effective method of detection of Assamese numerals captured under varied recording conditions and moods is presented here. The work also deals with gender variations taken as part of the speech recognition using a combination of Linear Predictive Code (LPC) and Vector Quantization applied to a combination of Self Organizing Map (SOM) and Multi Layer Perceptron (MLP) based system. The SOM and MLPs are used to constitute a Learning Vector Quantization (LVQ) block which is necessitated by the fact that the LPC-VQ methods fail to produce the expected outcome while dealing with numerals of Assamese- a language spoken by a sizable population in the North-Eastern part of India. The performance of the LPC feature set is further compared with that obtained PCA features applied to the same LVQ block. [ABSTRACT FROM AUTHOR]

Wet-and-low water changes of surface runoff have always been the focus of various researches. Regional water bodies are already being challenged by economic development. Artificial Neural Networks (ANN) deal with information through interactions among neurons (or nodes), approximating the mapping between inputs and outputs based on non-linear functional composition. They have the advantages of self-learning, self-organizing, and self-adapting. It is practical to use ANN technology to carry out forecast of surface runoff. [ABSTRACT FROM AUTHOR]

We investigate which of two Artificial Intelligence techniques is superior at making predictions about complex carcinogen systems. Artificial Neural Networks are shown to provide good predictions of carcinogen toxicology bands for drugs which are themselves used to treat cancerous cells, by using a novel system of molecular descriptors derived from the molecules΄ mass spectrometry intensities, reduced in dimensionality by Principal Component Analysis, to form a series of orthogonal descriptors which retain 95% of the variance of the original data. The creation of molecular descriptors from PCA-resolved mass spectrometry data is shown to be superior to the use of Self-Organising Maps, the selection of a series of modal fragments, or the use of every peak (within the confines of the precepts of Artificial Intelligence). A new system of backpropagation which increases network efficacy in this case is also proposed. [ABSTRACT FROM AUTHOR]

This study applies two empirical approaches such as logistic regression, and artificial network (ANN) to combine Conversion of Land Use and its Effects (CLUE-s) model to predict urban sprawl in the Paochiao watershed in Taipei County, Taiwan. The current investigation projected land-use dynamics for the next twenty years using demand prediction models such as the Markov chain and cellular automata model. Accordingly, the empirical models integrating with CLUE-s then simulated land-use patterns for future periods based on the predicted demands. Results show that land-use patterns predicted by the model tended to fragment and intersperse future patterns. The predicted land-use patterns have significant impact on urban sprawl in the Paochiao watershed. The study is an effective means of enhancing land-use monitoring and managing urbanized watersheds. [ABSTRACT FROM AUTHOR]

Due to the complex interaction of components, the design of concrete mix becomes difficult. This paper presents an artificial neural network (ANN) based visualization method to optimize the concrete mix design. It aims to minimize the cost of concrete such that all desired qualities are maintained. The procedure can be described as mapping data of concrete mix from multidimensional space to a two-dimensional plane with an ANN model, and then generating concrete property contours on this plane. The optimized mix proportions region can be determined intuitively based on the contours distribution. By means of an inversion mapping algorithm, the optimal point in this region can be mapped inversely to the original multidimensional space. Practical production test results show that good concrete mixes, which agree with the concrete compressive strength criterion and have lower cost, can be obtained. Application of this method can contribute significant benefits to the commercial concrete companies. [ABSTRACT FROM AUTHOR]

With the aim to study the faults diagnosis ability of the BPNN and the RBFNN, many experiments are done to test the learning ability, the diagnosis ability and the anti-noise ability. The analysis shows the RBFNN has better learning ability and anti-noise ability than the BPNN. However, in the process of concurrent faults diagnosis, both have bad recognition rate. A realistic application verifies the single neural network can not used for metallurgic fan machinery faults diagnosis. [ABSTRACT FROM AUTHOR]

Detection of subtle lesions on computed tomography (CT) images is a difficult task for radiologists, because subtle lesions such as small lung nodules tend to be low in contrast, and a large number of CT images must be interpreted in a limited time. One of the solutions to the problem, a temporal subtraction technique, has been introduced in the medical field as 2D visual screening. The temporal subtraction image is obtained by subtraction of a previous image from a current one, and can be used for enhancing interval changes on medical images by removing most of the normal background structures. In this study, we have developed a method for computerized detection of lung nodules by using temporal subtraction with a voxel-matching technique in multidetector-row CT (MDCT) images. First, the candidates for nodules were detected by use of a multiple threshold technique based on the pixel value in the temporal subtraction image obtained by the voxel-matching technique. Next, a number of features were quantified, and some false positives were removed by a rule-based method with an artificial neural network. We applied our computerized scheme to 6 chest MDCT cases including 94 lung nodules. Our scheme for detecting lung nodules provided a sensitivity of 71.2 % for lung nodules with sizes under 20 mm, with 9.8 and 11.5 false positives per scan on the consistency test and validation test, respectively. [ABSTRACT FROM AUTHOR]

In this study, a novel combined forecasting model integrating generalized linear auto-regression (GLAR) with artificial neural networks (ANN) was proposed to obtain accurate prediction results and ameliorate forecasting performances. We compare the performance of the new model with the two individual forecasting models-GLAR and ANN. Empirical results obtained reveal that the prediction using the combined model is generally better than those obtained using the individual model of GLAR and ANN in terms of the same evaluation measurements. Our findings reveal that the combined model proposed here can be used as an alternative forecasting tool for coal demand to achieve greater forecasting accuracy and improve prediction quality further. [ABSTRACT FROM AUTHOR]

Previous work in risk stratification of critically injured patients involved artificial neural networks (ANNs) of various configurations tuned to process traditional vital signs and demographical, clinical, and laboratory data obtained via direct contact with the patient. We now report ˵new vital signs″ (NVSs) that are superior in distinguishing the injured and can be derived without hands-on patient contact. Data from 262 trauma patients are presented, in whom NVSs derived from electrocardiogram (EKG) analysis (heart-rate complexity and variability) were input into a commercially available ANN. The endpoint was performance of life-saving interventions (LSIs) such as intubation, cardiopulmonary resuscitation, chest-tube placement, needle chest decompression, and blood transfusion. We conclude that based on EKG-derived NVS alone, it is possible to accurately identify trauma patients who undergo LSIs. Our approach may permit development of a next-generation decision support system. [ABSTRACT FROM AUTHOR]

It has been shown that generative representations, which allow the reuse of code, perform well on problems with high regularity (i.e. where a phenotypic motif must be repeated many times). To date, however, generative representations have not been tested on irregular problems. It is unknown how they will fare on problems with intermediate and low amounts of regularity. This paper compares a generative representation to a direct representation on problems that range from having multiple types of regularity to one that is completely irregular. As the regularity of the problem decreases, the performance of the generative representation degrades to, and then underperforms, the direct encoding. The degradation is not linear, however, yet tends to be consistent for different types of problem regularity. Furthermore, if the regularity of each type is sufficiently high, the generative encoding can simultaneously exploit different types of regularities. [ABSTRACT FROM AUTHOR]

The implementation of the inversion based linearizing method for the excitation system usually requires the full state feedback. This severely restricts its practical use because some of these states are very difficult to measure by physical sensors in practice due to no sensor available in the market or its high cost. To address this issue, the combined-inversion is proposed by coupling the left-inversion used as a soft-sensor with the right-inversion used as a nonlinear controller. Furthermore, to overcome the difficulty in explicit implementation of the combined-inversion by analytic means, an artificial neural network (ANN) is used to approximate the combined-inversion, thus forming the ANN combined-inversion. The simulation results demonstrate the effectiveness of the ANN combined-inversion. [ABSTRACT FROM AUTHOR]

The goal of speech event detection (SED) is to reveal the presence of important elements in the speech signal for different sound classes. In a speech recognition system, events can be combined to detect phones, words or sentences, or to identify landmarks with which a decoder could be synchronized. In this paper, we introduce three popular classification techniques, HMM, SVM, ANN and Non-Negative Matrix Deconvolution (NMD) for SED. The main purpose of this paper is to compare the performance of (1) HMM, (2) hybrid SVM/NMD (3) hybrid SVM/HMM and (4) hybrid MLP /HMM approaches to SED and emphasize approaches to reaching lower Event Error Rates (EER). It was found that the hybrid SVM/HMM approach outperformed the HMM system. Regarding EER, an improvement of 6% was achieved. The hybrid MLP/HMM got the best EER rate. Improvements of 11% and 8% were found in comparison with the HMM and hybrid SVM/HMM event detector, respectively. [ABSTRACT FROM AUTHOR]

To avoid the deficiencies which exist in the single-phase circuit for detecting harmonics and reactive currents, this paper presents a new detecting method after analysing the operating principle of the single-phase shunt active power filter(SAPF). This new detecting method omits the complex coordinate transform of traditional ip-iq detecting method, which can easily obtain sine and cosine signal that in phase of the power supply voltage that do not need PLL circuit. Meanwhile, this new detecting method also proposes the PID negative feedback which can quicken the dynamic response speed. At the same time, this paper also presents a new control method for harmonic and reactive power compensation with an artificial neural network (ANN) controller and a new control algorithm to eliminate harmonics, which makes actual output current of active power filter can accurately track the change of instruction current. Finally, this paper designs a SAPF that compensate harmonics and reactive currents for the single-phase circuit. Through simulation study with Matlab, the simulation results illustrate that the SAPF has a good performance in harmonic suppression and reactive power compensation. [ABSTRACT FROM AUTHOR]

In this paper, a novel approach to Arabic letter recognition is proposed. The system is based on the classified vector quantization (CVQ) technique employing the minimum distance classifier. To prove the robustness of the CVQ system, its performance is compared to that of a standard artificial neural network (ANN)-based solution. In the CVQ system, each input letter is mapped to its class using the minimum Euclidean distance. Simulation results are provided and show that the CVQ system always produces a lower Mean Squared Error (MSE) and higher success rates than the current ANN solutions. [ABSTRACT FROM AUTHOR]

The sedimentation characteristics in the Maasmond area of the approach channel of the port area of Rotterdam are studied using an Artificial Neural Network (ANN) technique. The important factors affecting the sedimentation process such as waves, wind, tides, surge, river discharge, etc. were studied, the corresponding time series data was analysed and the important variables behind the process were identified. An ANN model was built to predict sedimentation. The model was trained on the data collected during 1992-1998 and verified by the data of 1999-2000. The predictive accuracy of the ANN model was found to be adequate for the potential use in the operational decision making. [ABSTRACT FROM AUTHOR]