Group Method Of Data Handling Model Research Articles

Response model based on weighted bagging GMDH

Identifying customers who are more likely to respond to a mail or call is an important issue in direct marketing, and constructing response model is an effective approach to solve this problem. Logistic regression and neural network are two of the most commonly used models in direct marketing. However, both models have their own limitations. Based on the group method of data handling (GMDH) theory, in this study we propose a weighted bagging GMDH model (WB-GMDH) as a response model. First, multiple GMDH models are trained with bootstrap replicas of the original data set. Then, the final classification results are produced by aggregating the outputs of the base GMDH models based on the weights, which are computed according to the classification accuracy of each base model in the validation data set. The model is applied to two real-world data sets and its performance is measured by accuracy, ROC curve and lift chart. The results show that the WB-GMDH model has better performance than other models. Finally, analyses are carried out according to the explicit expression obtained from this model, which will help marketing analysts to develop new business plans better.

Prediction of partition coefficients of alkaloids in ionic liquids based aqueous biphasic systems using hybrid group method of data handling (GMDH) neural network

In this study, a hybrid GMDH–neural network model was developed in order to predict partition coefficients of alkaloids in aqueous biphasic system using different ionic liquids with the same inorganic salt. In order to accomplish this modeling, feed's weight percent compositions along with slope of tie-line (STL), tie-line length (TLL) and difference of molecular weight of salt and Ionic liquid were taken as the inputs and the desired partition coefficients (K) were estimated. Furthermore, the data set was divided into two parts: 80% of the data points were used for training and 20% for testing. For evaluation of the model's performance, partition coefficients obtained from the GMDH model were compared with their experimental values using different statistical measures. The proposed model can successively correlate and predict K-values and result a great agreement with the experimental data.

Integration of dynamic rainfall data with environmental factors to forecast debris flow using an improved GMDH model

The objective of this study was to apply an improved Group Method of Data Handling (GMDH) network model for prediction of debris flow by integrating dynamic rainfall data and environmental factors. The rainfall data were collected from weather information, and the environmental data were extracted from RS, GIS, drilling data, and geophysical data. The input variables used in the SAGA-GMDH model were derived from six variables acquired by Kernel Linear Discriminant Analysis (KLDA). The results showed that the GMDH for prediction of debris flow performed well using the training, validation, and testing sets (R2 above 0.80 and ARE below 3.54%). The SAGA-GMDH model was subsequently compared with a back-propagation (BP) neural network model and adaptive network fuzzy interference system (ANFIS). The accuracies of the SAGA-GMDH model prediction were slightly better than those of other two models, which demonstrated that the SAGA-GMDH model was more suitable for prediction of debris flow.

Source identification and characterization of atmospheric polycyclic aromatic hydrocarbons along the southwestern coastal area of Taiwan – with a GMDH approach

Air samples of polycyclic aromatic hydrocarbons (PAHs) were collected from May 2008 to April 2009 in the Kaohsiung coastal area of Taiwan to establish a dataset. Sample analysis suggested that PAH composition (dominated by 3-ring PAHs) and seasonal patterns (high concentration in winter and low concentration in summer) are nearly identical to those previously reported (Lai et al., 2011), except for a case of road construction during the sample period. The predominant sources of PAHs were identified by isomer ratio, HCA and PCA analyses as vehicle emissions and coal/wood combustion. According to the results of parameters selection via the Group Method of Data Handling (GMDH), the variations in gaseous PAHs and particulate PAHs were related mainly to meteorological conditions and to routine monitoring of air pollutant concentrations, respectively. A comparison of model prediction accuracy between particulate and gaseous PAHs suggests better prediction for particulate PAHs than gaseous PAHs, due to dramatic variation in meteorological conditions which results in higher prediction uncertainty for gaseous PAHs. Model predictions of PAH concentrations showed satisfactory agreement with measured data, except for specific incidents, such as joss paper burning events. The results of this study suggest that the GMDH model may be applied to PAH concentration prediction for the study area and may be applicable for other countries based on routine monitoring of air pollutant concentrations and meteorological conditions.

Artificial neural networks based prediction of performance and exhaust emissions in direct injection engine using castor oil biodiesel-diesel blends

In this study, the performance and emission characteristics of a direct injection diesel engine using castor oil biodiesel (COB)-diesel blended fuels were investigated experimentally and then predicted by artificial neural networks. For this aim, castor oil was converted to its biodiesel via transesterification approach. Then, the effects of the biodiesel percentage in blend, engine load, and speed on brake power, brake specific fuel consumption (BSFC), nitrogen oxides (NOx), carbon dioxide (CO2), carbon monoxide (CO), and particle matter (PM) have been considered. Fuel blends with various percentages of biodiesel (0%, 5%, 10%, 15%, 20%, 25%, and 30%) at various engine speeds and loads were tested. The results indicated that blends of COB with diesel fuel provide admissible engine performance; on the other side, emissions decreased so much. Two types of neural networks, a group method of data handling (GMDH) and feed forward were used for modeling of the diesel engine to predict brake power, BSFC, and exhaust emissions such as CO, CO2, NOx, and PM values. The comparison results demonstrate the superiority of the feed forward neural network models over GMDH type models in terms of the statistical measures in the training and testing data but in the number of hidden neurons and model simplicity, GMDH models are preferred.

MONITORING SEVERE ACCIDENTS USING AI TECHNIQUES

After the Fukushima nuclear accident in 2011, there has been increasing concern regarding severe accidents in nuclear facilities. Severe accident scenarios are difficult for operators to monitor and identify. Therefore, accurate prediction of a severe accident is important in order to manage it appropriately in the unfavorable conditions. In this study, artificial intelligence (AI) techniques, such as support vector classification (SVC), probabilistic neural network (PNN), group method of data handling (GMDH), and fuzzy neural network (FNN), were used to monitor the major transient scenarios of a severe accident caused by three different initiating events, the hot-leg loss of coolant accident (LOCA), the cold-leg LOCA, and the steam generator tube rupture in pressurized water reactors (PWRs). The SVC and PNN models were used for the event classification. The GMDH and FNN models were employed to accurately predict the important timing representing severe accident scenarios. In addition, in order to verify the proposed algorithm, data from a number of numerical simulations were required in order to train the AI techniques due to the shortage of real LOCA data. The data was acquired by performing simulations using the MAAP4 code. The prediction accuracy of the three types of initiating events was sufficiently high to predict severe accident scenarios. Therefore, the AI techniques can be applied successfully in the identification and monitoring of severe accident scenarios in real PWRs.

River flow time series using least squares support vector machines

Abstract. This paper proposes a novel hybrid forecasting model known as GLSSVM, which combines the group method of data handling (GMDH) and the least squares support vector machine (LSSVM). The GMDH is used to determine the useful input variables which work as the time series forecasting for the LSSVM model. Monthly river flow data from two stations, the Selangor and Bernam rivers in Selangor state of Peninsular Malaysia were taken into consideration in the development of this hybrid model. The performance of this model was compared with the conventional artificial neural network (ANN) models, Autoregressive Integrated Moving Average (ARIMA), GMDH and LSSVM models using the long term observations of monthly river flow discharge. The root mean square error (RMSE) and coefficient of correlation (R) are used to evaluate the models' performances. In both cases, the new hybrid model has been found to provide more accurate flow forecasts compared to the other models. The results of the comparison indicate that the new hybrid model is a useful tool and a promising new method for river flow forecasting.

Self-Organizing Polynomial Networks for Time-Constrained Applications

A metamodeling of complex calculation procedures (static systems) is investigated with the aim to create low-complexity surrogate models that are applicable in low-computing-power real-time (RT) measurement systems. Unlike the single-parameter error criteria and minimum description length measure, the proposed compound-squared-relative-error measure forces the group method of data handling (GMDH) algorithm to prefer the models having the smallest compound deviation of accuracy and execution time from the given thresholds and thus generally leads to more favorable models with respect to both conditions. Approximation errors, execution speed, and the applicability of the derived GMDH models in RT flow-rate measurements of natural gas are discussed and compared with the corresponding models derived by artificial neural network and support vector regression.

The Prediction Of Air Pollution By Using Neuro-fuzzy Gmdh

Air pollution is one of today modern life phenomena. It is resulted to create round-the-clock human begin activities. Control of environment pollution is complicated scientific process that enters policy, economy, technology and sociology. GMDH (Group Method of Data Handling) has been used for the identification of a mathematical model that has many input variables but limited data needs by using a hierarchical structure. This paper proposes a Neuro-fuzzy GMDH model, adopting Gaussian radial basis functions (GRBF) as partial descriptions of GMDH. GRBF is reinterpreted as both a simplified fuzzy reasoning model and as a three-layered neural network. In this paper, is used Neuro-fuzzy GMDH algorithm for predicting air pollution data and then were compared the results of predicting air pollution data by using Neuro-fuzzy GMDH and Multi Layer Perceptron (MLP).

Diagnostics of Loss of Coolant Accidents Using SVC and GMDH Models

As a means of effectively managing severe accidents at nuclear power plants, it is important to identify and diagnose accident initiating events within a short time interval after the accidents by observing the major measured signals. The main objective of this study was to diagnose loss of coolant accidents (LOCAs) using artificial intelligence techniques, such as SVC (support vector classification) and GMDH (group method of data handling). In this study, the methodologies of SVC and GMDH models were utilized to discover the break location and estimate the break size of the LOCA, respectively. The 300 accident simulation data (based on MAAP4) were used to develop the SVC and GMDH models, and the 33 test data sets were used to independently confirm whether or not the SVC and GMDH models work well. The measured signals from the reactor coolant system, steam generators, and containment at a nuclear power plant were used as inputs to the models, and the 60 sec time-integrated values of the input signals were used as inputs into the SVC and GMDH models. The simulation results confirmed that the proposed SVC model can identify the break location and the proposed GMDH models can estimate the break size accurately. In addition, even if the measurement errors exist and safety systems actuate, the proposed SVC and GMDH models can discover the break locations without a misclassification and accurately estimate the break size.

A HYBRID GMDH AND LEAST SQUARES SUPPORT VECTOR MACHINES IN TIME SERIES FORECASTING

Time series consists of complex nonlinear and chaotic patterns that are difficult to forecast. This paper proposes a novel hybrid forecasting model which combines the group method of data handling (GMDH) and the least squares support vector machine (LSSVM), known as GLSSVM. The GMDH is used to determine the useful input variables for the LSSVM model and the LSSVM model that works as time series forecasting. Three well-known time series data sets are used in this study to demonstrate the effectiveness of the forecasting model. These data are utilized to forecast through an application aimed to handle real life time series. The results found by the proposed model were compared with the results of the GMDH and LSSVM models. Experiment result indicates that the hybrid model was a powerful tool to model time series data and provides a promising technique in time series forecasting methods.

Smart Soft-Sensing for the Feedwater Flowrate at PWRs Using a GMDH Algorithm

The thermal reactor power in pressurized water reactors (PWRs) is typically assessed using secondary system calorimetric calculations based on accurate measurements of the feedwater flowrate. Therefore, precise measurements of the feedwater flowrate are essential. In most PWRs, Venturi meters are used to measure the feedwater flowrate. However, the fouling phenomena of the Venturi meter deteriorate the accuracy of the existing hardware sensors. Consequently, it is essential to resolve the inaccurate measurements of the feedwater flowrate. In this study, in order to estimate the feedwater flowrate online with high precision, a smart soft sensing model for monitoring the feedwater flowrate was developed using a group method of data handling (GMDH) algorithm combined with a sequential probability ratio test (SPRT). The uncertainty of the GMDH model was also analyzed. The proposed sensing and monitoring algorithm was verified using the acquired real plant data from Yonggwang Nuclear Power Plant Unit 3.

Estimation of Liquid−Liquid Equilibrium for a Quaternary System Using the GMDH Algorithm

Liquid−liquid equilibrium (LLE) data are important for designing and modeling of process equipment. In this article, the group method of data handling (GMDH) technique has been used to estimate LLE data of the quaternary system of corn oil + oleic acid + ethanol + water at 298.15 K. Using this method, a new model has been proposed that is suitable for use in predicting LLE data. The equilibrium data presented herein were predicted using the GMDH model. To evaluate the performance of the GMDH method, we compared the data predicted using the proposed model with experimental data. Moreover, we compared the deviation between the experimental data and the data estimated with the GMDH. We also used the deviations of some existing models for the comparison. The results show that the proposed model can be used as an efficient tool for estimating LLE data.

A GMDH neural network-based approach to passive robust fault detection using a constraint satisfaction backward test

This paper proposes a new passive robust fault detection scheme using non-linear models that include parameter uncertainty. The non-linear model considered here is described by a group method of data handling (GMDH) neural network. The problem of passive robust fault detection using models including parameter uncertainty has been mainly addressed by checking if the measured behaviour is inside the region of possible behaviours based on the so-called forward test since it bounds the direct image of an interval function. The main contribution of this paper is to propose a new backward test, based on the inverse image of an interval function, that allows checking if there exists a parameter in the uncertain parameter set that is consistent with the measured system behaviour. This test is implemented using interval constraint satisfaction algorithms which can perform efficiently in deciding if the measured system state is consistent with the GMDH model and its associated uncertainty. Finally, this approach is tested on the servoactuator being a FDI benchmark in the European Project DAMADICS.

Multiple GMDH models for estimating resource requirements

Accurate estimation or prediction of the resource required for a project is very important for construction. The more accurate the prediction model, the greater the potential for cost savings will be through elimination of any redesign and the minimization of the maintenance expenses. Contractors can also make use of the models for last‐minute bid estimation. In the past the estimators perform the task by analogy with similar previous projects. This approach highly relies on their experience and knowledge. Owing to the lack of a scientific and easily apprehensible method in resource estimation, prediction outcomes are mainly based on humans’ perception, which is inconsistent and exhibits large variations. This paper proposes the use of multiple Group Method of Data Handling (GMDH) models in developing models for resource estimation. The illustrative example has demonstrated the high accuracy of the approach which is superior to other architectures based on artificial neural networks.

Monitoring an Experimental Reactor Using the Group Method of Data Handling Approach

This paper presents a reactor-monitoring algorithm using the group method of data handling (GMDH) that creates nonlinear algebraic models for system characterization. The monitoring system was applied to the IEA-R1 experimental reactor at the Instituto de Pesquisas Energéticas e Nucleares (IPEN). The IEA-R1 is a 5-MW pool-type research reactor that uses light water as coolant and moderator and graphite as reflector. The GMDH provides a general framework for characterizing the relationships among a set of state variables of a process system and is used for generating estimates of critical variables in an optimal data-driven model form. The monitoring system developed in this work was used to predict the IEA-R1 reactor environment, using nuclear power, rod position, and coolant temperatures, by combining two variables at a time. The results obtained using the GMDH models agreed very well with the dose rate measurements, with prediction errors of less than 5%. The error was minimal when the dose rate prediction was made using reactor power and coolant temperature.

Modelling hook times of mobile cranes using artificial neural networks

The hook times of mobile cranes are processes that are of non‐linear and discrete nature. Artificial neural network is a data processing technique that lends itself to this kind of problem. Three common artificial neural network architectures – multi‐layer feed‐forward (MLFF), group method of data handling (GMDH) and general regression neural network (GRNN) – are compared. The results show that the GRNN model aided with genetic algorithm (GA) is most promising in describing the non‐linear and discrete nature of the hook times. The MLFF model can also give a moderate level of accuracy in the estimation of hook travelling times of mobile cranes and is ranked second. The GMDH model is outperformed by the former two due to a less promising R‐square.

A self-organization algorithm for real-time flood forecast

The group method of data handling (GMDH) algorithm presented by A. C. Ivakhnenko and colleagues is an heuristic self-organization method. It establishes the input–output relationship of a complex system using a multilayered perception-type structure that is similar to a feed-forward multilayer neural network. This study provides a step towards understanding and evaluating a role for GMDH in the investigation of the complex rainfall–runoff processes in a heterogeneous watershed in Taiwan. Two versions of the revised GMDH model are implemented: a stepwise regression procedure and a recursive formula. Eleven typhoon events in the Shen-cei Creek watershed, Taiwan, are used to build the model and verify its usefulness. The prediction results of the revised GMDH models and the instantaneous unit hydrograph (IUH) model are compared. Based on the criteria of forecasting precision and the rate and time of peak error, a much better performance is obtained with the revised GMDH models. Copyright © 1999 John Wiley & Sons, Ltd.

Neuro-fuzzy GMDH and its application to modelling grinding characteristics

Mathematical models, in which many input variables are involved, require a range of input and output data, since the number of parameters increases with the input variables. GMDH (Group Method of Data Handling) has been used for the identification of a mathematical model that has many input variables but limited data needs by using a hierarchical structure. This paper proposes a neuro-fuzzy GMDH model, adopting Gaussian radial basis functions (GRBF) as partial descriptions of GMDH. GRBF is reinterpreted as both a simplified fuzzy reasoning model and as a three-layered neural network. As an example of applying the algorithm, the wheel wear equation is identified, using data from experiments of abrasive cut-off. In the model, characteristics of work materials, grinding fluids, factors of wheels, wheel velocity and table feed are used as input variables, and the grinding ratio is the resulting output. The validity of the model is confirmed within the predicted accuracy by using additional data.

GMDH algorithm as a tool for bivalve growth analysis and prediction

The question of whether growth in bivalves is predictable in terms of environmental condition is addressed directly by trying to infer juvenile scallop growth from environment data within and between two locations in the Baie des Chaleurs, Québec. Using models based on either self-organizing models – the group method of data handling (GMDH) algorithm - or multilinear regressions, scallop growth was found to be predictable. GMDH models lead consistently to better predictions than multilinear regressions and could thus be a useful alternative tool in managing scallop fisheries aquaculture. Temperature and food availability were the most prominent variables included in the GMDHD models, emphisizing their importance as physical determinants of scallop growth.