The artificialneural networks (ANN) technologies provide on-line capability to analyze many inputs and provide information to multiple outputs, and have the capability to learn or adapt to changing conditions. No doubt that the determination ofSkin permeability is a time consuming process; which involves a quite tedious work. Material and method: Software Neurodimension was used for this study. A data set was taken from literature and used to train the network. A set of 20 compounds were used to construct the ANN models for training and 10 compounds used for predictionofskinpenetration (n=30, molecular weight>500 da). Skin permeability expressed in log Kp (cm/h). Abraham descriptors of R 2 (excess molar refraction), π2 H (dipolarity/polarizability), Σα 2 H, Σβ 2 H (the overall or
Abstract. The masonry is not only included among the oldest building mate- rials, but it is also the most widely used material due to its simple construction and low cost compared to the other modern building materials. Nevertheless, there is not yet a robust quantitative method, available in the literature, which can reliably predict its strength, based on the geometrical and mechanical characteristics of its components. This limitation is due to the highly nonlinear relation between the compressive strength of masonry and the geometrical and mechanical properties of the components of the masonry. In this paper, the application of arti ﬁcial neural networks for predicting the compressive strength of masonry has been investigated. Speciﬁcally, back-propagation neuralnetwork models have been used for predicting the compressive strength of masonry prism based on experimental data available in the literature. The comparison of the derived results with the experimental ﬁndings demonstrates the ability of arti ﬁcial neural networks to approximate the compressive strength of masonry walls in a reliable and robust manner.
face milling with resultant cutting force and selected machining parameters and demonstrated its adequacy. Thamizhmnaii et al. (2008) reported that higher flank wear occurred in low cutting speed with high feed rate and depth of cut in turning of SS 440 C stainless steel. A spanking new Transductive-Weighted Neuro-Fuzzy Inference Technique (TWNFIS) was proposed (Agustin et al., 2009) to model tool wear in turning and proved the accuracy by comparing with experimental values. Li et al. (2002) used vibration signals to find out drill wear and proposed a relationship between the vibration and the tool wear with fuzzy neuralnetwork model. It was also demonstrated that features of vibration signals can be used to determine the drill wear with greater accuracy. Tansel et al. (2000) proposed cutting force wear relation by Force-variation-based encoding (FVBE) and Segmental-average-based encoding (SABE) methods and proved both are excellent performance in wear estimation. Neuralnetwork was applied to predict the wear during hard turning of AISI H-13 steel and the proposed model provided better prediction capabilities (Tugrul and Yigit, 2005). An efficient and successful relationship was established by Choudhury and Bartarya (2003) between tool wear and surface roughness along with cutting temperature. An online tool wear monitoring technique was developed (Silva et al., 1998) with input signals as cutting force, spindle current, sound and vibration in turning and demonstrated efficiency of the suggested technique. Jurkovic et al. (2005) presented direct tool wear measurement methodology using machine vision with 3D picture of tool relief surface. An analytical computation of flank wear was expressed and predicted wear at various cutting speeds was compared with experimental values (Bouzid, 2005). Palanisamy et al. (2008) suggested regression analysis and artificialneuralnetwork model to forecast the flank wear which were validated through experiments.
Artificialneural networks (ANNs), as one of the most attractive branches in artificial intelligence, has the potential to handle problems such as modeling, estimating, prediction, optimization, diagnosis, and adaptive control in complex nonlinear systems .The capabilities of ANNs in capturing the mathematical mapping between input variables and output features are of primary significance for modeling machining processes . An ANN is essentially a mathematical model that mimics the human reasoning and neurobiology and is based on the following assumptions: information processing occurs in a number of simple elements called neurons; signals are transmitted between neurons over connection links; each connection link has an associated weight that multiplies the signal transmitted; each neuron applies an activation function to the incoming signal to determine its output signal. Multi layer feedforward network have been applied successfully to solve some difficult and diverse problems by training them in a supervised manner with a highly popular algorithm known as the error back propagation algorithm. This algorithm is based on error correction learning rule. Figure 1 illustrates a schematic structure of proposed multilayer feedforward backpropagation neuralnetwork. The backward linkages are used only for the learning phase, whereas the forward connections are used for both the learning and operational phases.
Abstract—This study aims to improve water level prediction at Bedup River with estimations made to absent precipitation data, both usingArtificialNeuralNetwork (ANN). Studies to predict water level in the state of Sarawak, Malaysia have been actively carried out. However, among problem faced was absent precipitation readings, which inevitably affected water level precipitation accuracies. Backpropagation properties of ANN was used in the study to predict both missing precipitation and water level. ANN model developed in this study successfully estimates missing precipitation data of a recorder in Bedup River, Sarawak with 96.4% accuracy. The predicted values of precipitation were then used to forecast water level of the same gauging station and yielded accuracy value of 85.3%, compared to only 71.1% accuracy of water level prediction with no estimation made to its missing precipitation data. These results show that ANN is an effective tool in forecasting both missing precipitation and water level data, which are utmost essential to hydrologists around the globe.
Abstract—Compositions measurement is a vitally critical issue for the modelling and control of distillation process. The product compositions of distillation columns are traditionally measured using indirect techniques via inferring tray compositions from its temperature or by using an online analyser. These techniques were reported as inefficient and relatively slow methods. In this paper, an alternative procedure is presented to predict the compositions of a binary distillation column. Particle swarm opti- misation based artificialneuralnetwork PSO-ANN is trained by different algorithms and tested by new unseen data to check the generality of the proposed method. Particle swarm optimisation is utilised, here, to choose the optimal topology of the network. The simulation results have indicated a reasonable accuracy ofprediction with a minimal error between the predicted and simulated data of the column.
finding an appropriate neuralnetwork, able to simulate the behavior of proximity sensor at different functioning conditions, we choose a high degree of parameterization for the application, with different kind of parameters, which varies in a large range. The second step is to determine the neuralnetwork , based on over 40,000 data sets obtained from multiparametric solving of the numerical model and going through a selection process based on network performance criteria. With network determined in this way, simulations were made on the data sets not used for the training process.
In transformer incipient fault recognition, the relationship between gas type and fault is non- linear. This causes problem in the convergence rate and oscillation in artificialneuralnetwork (ANN). The parameters of the ANN must also be properly tuned in order to obtain the best performance of the network. To overcome these problems, in this work, a method of combina- tion ofartificialneuralnetwork (ANN) and various particle swarm optimisation (PSO) tech- niques to predict transformer incipient fault has been successfully proposed. In this method, the ANN was used to identify the transformer incipient fault and various techniques of PSO were applied to optimise the performance of the ANN. The performance of various PSO tech- niques in combination with ANN was compared with the existing DGA method, ANN alone and previously reported work to identify the best method for transformer incipient fault
ArtificialNeuralNetwork (ANN) can be useful tool to predict the hydrodynamic coefficients of permeable breakwater. The stability and reliability analyses of coastal structures such as rubble mound breakwaters usingneural networks have been carried out [Mase et al, (1995)]; [Kim and Park, (2005)]. Similarly [Mandal et al., (2009)] have applied neuralnetwork to predict the wave transmission of the floating breakwater. The current study aims to develop quick and easy reliable method to predict the hydrodynamic coefficients; wave transmission and wave reflection of permeable paneled breakwater usingArtificialNeural Networks.
Electric load demand is a function of weather variables and human social activities, industrial activities as well as community developmental level to mention a few [2-7]. Statistical techniques and Expert system techniques have failed to adequately address this issue [2-10]. The daily operation and planning activities of an electric utility requires the predictionof electricity demand of its customers. In general, the required load forecasts can be categorized into short-term, mid-term, and long-term forecasts. The short-term forecasts refer to hourly predictionof the load for a lead time ranging from one hour to several days out. The mid-term forecasts can either be hourly or peak load forecasts for a forecast horizon of one to several months ahead. Scheduling of fuel purchases, load flow studies or contingency analysis, and planning for energy, while the long-term forecasts refer to forecasts made for one to several years in the future. The quality of short-term hourly load forecasts has a significant impact on the economic operation of the electric utility since decisions such as economic scheduling of generating capacity, transactions such as ATC (Available Transmission Capacity) are based on these forecasts and they have significant economic consequences.
Abstract: - An artificialneuralnetwork model for the predictionof path loss in urban macrocellular environment is presented. The model consists of a multilayer perceptron trained with measured data using Scaled Conjugate Gradient algorithm. Comparison between the proposed model on one hand, and the free space, Hata and Egli models on the other hand shows a better prediction result. With the proposed ANN model a good generalization is achieved, and it is accurate in environments different from the one used in training the network.
The method developed in this study makes it possible to examine the effect of different types of ammunition on the barrel using computers and eliminates the need for time consuming and costly tests. In addition, by integrating an artificialneuralnetwork trained according to barrel character- istics to the software, which is controlling barrel position, the shooting accuracy and strike power of the weapon system can be increased by simply adjusting the initial position of the barrel. This would make it possible to design weapons that are lighter and more effective against targets. The velocity of a projectile inside the barrel varies by time and forces the barrel to change its natural frequencies continuously. This means that for different projectiles and muzzle velocities, different vibration modes are created in the barrel. For example, the muzzle displacement value is positive at some muzzle velocities, and negative at others. In addition, predicting the amount of muzzle displacement in a weapon barrel may not be sufficient sometimes, predicting the angle of inclination of the barrel may also be required. The neuralnetwork modelled in this study does not require many complex systems to make prediction, but an artificialneuralnetwork with at least two hid- den layers is required only, and a preparation of a larger training set that represents the problem space are needed to predict both positive and negative muzzle displacements. Using the proposed method may help engineers in improving the target accuracy of a weapon system.
ABSTRACT: Artificialneural networks (ANN) are computational models inspired by the neural systems of living beings capable of learning from examples and using them to solve problems such as non-linear prediction, and pattern recognition, in addition to several other applications. In this study, ANN were used to predict the value of the area under the disease progress curve (AUDPC) for the tomato late blight pathosystem. The AUDPC is widely used by epidemiologic studies of polycyclic diseases, especially those regarding quantitative resistance of genotypes. However, a series of six evaluations over time is necessary to obtain the final area value for this pathosystem. This study aimed to investigate the utilization of ANN to construct an AUDPC in the tomato late blight pathosystem, using a reduced number of severity evaluations. For this, four in- dependent experiments were performed giving a total of 1836 plants infected with Phytophthora infestans pathogen. They were assessed every three days, comprised six opportunities and AUDPC calculations were performed by the conventional method. After the ANN were created it was possible to predict the AUDPC with correlations of 0.97 and 0.84 when compared to con- ventional methods, using 50 % and 67 % of the genotype evaluations, respectively. When using the ANN created in an experiment to predict the AUDPC of the other experiments the average correlation was 0.94, with two evaluations, 0.96, with three evaluations, between the predicted values of the ANN and they were observed in six evaluations. We present in this study a new paradigm for the use of AUDPC information in tomato experiments faced with P. infestans. This new proposed paradigm might be adapted to different pathosystems.
In all measurements three fruits at the stage of maturity (yellowing) were used from three plants per plot. The average fruit weight was determined on an electronic scale. Fruit length (longitudinal) and diameter (transversal) were measured with calipers. The flesh thickness was measured as lateral length of the fruit flesh, using a caliper. The fruit firmness was measured after dividing the fruit in half, in the transverse direction. The firmness was determined at four equidistant points on each fruit half, at a distance of 0.5 cm from the skin, based on resistance to penetrationof the flesh. For this purpose, a penetrometer (Fruit Pressure Tester, Italy; model 53205) with an adapter (height 3.0 cm, diameter 3.0 cm) was used.
some effects on how well the ANN trains. Here we use Scaled normalization to bring the data into dyna- mic range of the tangent sigmoid transfer function of the network. Before training the ANNs, the parame- ters ofnetwork including the number of nodes in the hidden layer, weights and biases learning rates and momentum values were optimized. In order to deter- mine the optimum number of nodes in hidden layer several training sessions were conducted with diffe- rent number of hidden nodes. The values of standard error of training (SET) and standard error ofprediction (SEP) were calculated after each 1000 iterations and calculation was stopped when overtraining began, then SET and SEP values were recorded. The recorded values of SET and SEP were plotted against the number of nodes in hidden layer, and the number of hidden nodes with minimum values of SET and SEP was chosen as the optimum one (Figure 3). It can be seen from this figure that 6 nodes in the hid- den layer were sufficient for a good performance of the network. Learning rates of weights and biases and also momentum values were optimized in a simi- lar way and the results are shown in Figures 4-6, res-
Cognitive radio is viewed as an intelligent way of utilizing the spectrum efficiently depending on the environment. Spectrum sensing serves to be the most important part in implementing a cognitive radio system. The method followed to sense the frequency spectrum must maintain accuracy. Wireless signals are subjected to many forms of distortion and disturbances depending on the factors like distance, transmission medium and so on. These types of distortions and noises tend to lower the ratio of signal to noise(SNR). So when the signal is sensed at any point, depending on the factors which has affected the signal, the signal exhibits low SNR. Cyclostationary feature detection process is one of the ways to detect the absence or presence ofa particular signal effectively even when the signal exhibits low SNR.Cyclostationary process is defined as a random processfor which statistical properties like mean, autocorrelation changes periodically with time . These processes are normally caused by modulation, coding or can be inserted for information recovery . The cyclostationary features in a particular signalalways exhibits regenerative periodicity which is considered as one of its characteristic property . There are mainly two types of methods used to realize cyclostationary feature detection.
The aim of this research is to quantify the tourism demand using an ArtificialNeuralNetwork (ANN) model. The methodology was focused in the treatment, analysis and modulation of the tourism time series: “Monthly Guest Nights in Hotels” in Northern Portugal recorded between January 1987 and December 2006, since it is one of the variables that better explain the effective tourism demand. The model used 4 neurons in the hidden layer with the logistic activation function and was trained using the Resilient Backpropagation algorithm. Each time series forecast depended on 12 preceding values. The developed model yielded acceptable goodness of fit and statistical properties and therefore it is adequate for the modulation and predictionof the reference time series.
Abstract: Problem statement: The purpose of this study is to develop an artificial immune system for recognizing stock market trends and predict upward and downward directions of stock market. This study compared two prediction models, an Artificial Immune System (AIS) and ArtificialNeuralNetwork (ANN) for predicting the future index value, trend of Indian stock market and discovers the best prediction model. Approach: AIS is an efficient system for predicting trend due to its high capability of learning and retaining information in memory. Our proposed system was tested using SENSEX (Sensitive Index) data from Bombay Stock Exchange (BSE) of India. Results: Performance of models have been evaluated on the basis of the simulation results done on MATLAB. Experiments have been performed for both methods on well-known technical indicators and compared their results with SENSEX data. Conclusion: Artificial Immune System is more efficient than ArtificialNeuralNetwork.
environment . Image matching is used to make decision for object . An adaptive neuralnetwork can accept visual signals as inputs directly from visual sensors for the spatial information . ANNs can plan motion of several mobile robots ensuring collision avoidance . Fuzzy modelling of the real robot's environment using Hopfield neuralnetwork has been used . ANN base motion planer can respond to changing real time situation . Latency effect in a closed loop system can be reduced for motion prediction . To design the vehicle controller behavioral cloning machine learning algorithm and neuralnetwork algorithms can also be used . Improvements have been made using fuzzy logic based path planning algorithm and are more effective. . Convergence of a NeuralNetwork has been improved using Q-Learning (NCQL) algorithm . Extended Back Propagation Algorithm predicts moving obstacles for obstacle avoidance . Designing a robot and planning its motion is always a new task in a new environment.
Khadim Moin Siddiqui received his B.Tech degree in Electronics Engineering from Azad Institute of Engineering & Technology, Lucknow, India, in 2007, the M.Tech degree in Power Electronics & Drives from Madan Mohan Malaviya Engineering College, Gorakhpur, India, in 2011. He has published 8 national as well as international research papers. He has received best paper award in the IEEE international conference organized by IEEE India chapter at Noorul Islam University Chennai, March, 2012. He was a teaching, lecturer with Department of Electronics Engineering, Azad Institute of Engineering & Technology, Lucknow in August 2008 to 2009. He worked as a trainee in NIIT, Bangalore, in 2007 to July 2008.At present, he is working as Assistant Professor in Jaipur National University (JNU), Jaipur. His research interests include power and control, Fault Diagnosis and identification of motor and on GSM voice synthesis.