adaptive neuro-fuzzy inference system

Analyzing financial ratios over consecutive years is beneficial for evaluating the financial performance of construction companies. However, such an analysis can be tedious due to the vast number of the ratios. Therefore, developing an expert system based on artificial intelligence algorithms to identify and predict factors influencing the construction companies' financial performance is essential. To this end, a hybrid model based on Genetic Algorithm (GA) and Adaptive Neuro-Fuzzy Inference System (ANFIS) was introduced in this research to predict the financial performance of construction companies in Iran. This research is applied as descriptive and in terms of methodology well developed; also conducted cross-sectionally. The statistical population included all active construction companies in the construction sector in Tehran. Due to time and resource constraints, a random sampling technique was used. A questionnaire was utilized for data collection and data analysis, factor analysis methods and neuro-fuzzy system combined with GA were employed. The ANFIS combined with GA can evaluate the construction companies' financial performance with the minimum error. The findings ultimately resulted development of a model that forecasts the financial performance of Iranian construction companies, allowing them to concentrate on factors that improve financial performance.

Seismic designs and numerical analyses required fundamental parameters such as damping ratio and shear modulus. In this study, large-scale triaxial cyclic tests were used to investigate the dynamic properties of limestone ballast and electric arc furnace (EAF) slag ballast. The term ‘shear stiffness’ is typically reported in a normalized form using shear modulus. As the laboratory test results showed, an increase in confining pressure, loading frequency and anisotropy raises the shear modulus of materials. Shear modulus and damping ratio values do not appear to be significantly affected by an increase in loading cycles. Loading frequency plays the most significant role in changing damping ratio values. An adaptive neuro-fuzzy inference system (ANFIS) was also used to predict the normalized shear modulus and the damping ratio in this study. The results of the developed model were consistent with those of the laboratory tests. Moreover, the relations among the dynamic properties were estimatedly determined using the nonlinear regression method.

Dynamic parameters are the most important geotechnical data used to understand the behavior of soil media under dynamic loads and to recognize the seismic response of the soil. Several in-situ and laboratory geophysical tests, such as the down-hole test, are used to determine these parameters. Since this experiment is costly and time-consuming and the preparation of appropriate boreholes is not easy, it is preferable to estimate the results of this test with the help of empirical correlations or experimental models. The main output of the down-hole test is the shear wave velocity (VS) of soils, which can be used to obtain the dynamic shear modulus (Gs) indirectly. The relationship between physical properties and mechanical specifications of soils is a well-known principle of geotechnical engineering. Utilizing the results of 19 down-hole experiments and available geotechnical data in the southern regions of Tehran, as well as the inputs of an adaptive neuro-fuzzy inference system (ANFIS). This study attempts to provide practical models to predict shear wave velocity of fine-grained soils in Tehran. Two new models have been proposed as a result of preprocessing and smart modeling. The independent variables of the first suggested model included the moisture content, plasticity index (PI), liquid limit (LL), depth of test, and grain size distribution of soils. In the second model, the number of standard penetration test (NSPT) is also used in addition to the mentioned independent variables. The proposed models had coefficients of determination (R2) of 0.74 and 0.8 for the total training and validation data, respectively.

In textile and garment industry, the physiological comfort of fabric as one of the important parameters, can be improved by the fabric finishing treatment. Nevertheless, the toxic chemicals produced in this process leads to the pollution of the environment. Therefore, this study aims to improve the physiological comfort of the cotton fabric without applying the finishing process as green technology. Accordingly, air permeability and moisture transfer as two important parameters of the fabric physiological comfort are evaluated with the structural parameters of the cotton yarn using experimental and theoretical procedures. For theoretical evaluations, a novel neuro-fuzzy network (ANFIS) is proposed and used for modelling and estimation. The structural parameters of yarn are the yarn linear density, yarn twist and fineness of fibers, which are defined as inputs and air permeability and moisture transfer of the cotton samples are considered as the outputs of developed ANFIS model. According to the experimental and modeling results, the fiber fineness, yarn linear density (Ne) and yarn twist have the same effect on the output parameters. It is also found that both parameters of the physiological comfort sensory can be improved effectively without finishing process. Simulation results show the novel proposed ANFIS that has high learning capability, fast convergence and accuracy greater than 99% and negligible error value smaller than of 1% can be reasonably used in textile industry. In addition, for the winter garments, the optimum points of turns per meter (T.P.M) coefficient, English count of yarn, and fibers fineness are 4.5, 25 and 3, respectively.

Punching shear capacity is a key factor for governing the collapsed form of slabs. This fragile failure that occurs at the slab-column connection is called punching shear failure and has been of concern for the engineers. The most common practice in evaluating the punching strength of the concrete slabs is to use the empirical expressions available in different building design codes. The estimation of punching loads involves experimental setup which is time-consuming, uneconomical and also, more manpower and materials are required. The present study demonstrates the use of data mining techniques as a substitute of former to predict the punching loads on the variation of various parameters. In this study, various type of data mining techniques including Adaptive Neuro-fuzzy Inference System (ANFIS), Artificial Neural Network (ANN) and Generalized Neural Network (GRNN) were applied to model and estimate the punching load of reinforced concrete slab–column connections. For the study, a data set consisting of 89 observations from available literature was analysed and randomly selected 62 observations were used for model development whereas the rest 27 were used to test the developed models. While the outcomes of ANN and GRNN model provides suitable estimation performance, the Gaussian membership based ANFIS model performed best in the determination of coefficient of correlation (Cc). Sensitivity study indicates that the parameter effective depth of slab (d) is the most influencing one for the estimation of punching load of reinforced concrete slab–column connections for this data set.

In large open pit mines prediction of Peak Particle Velocity (PPV) provides useful information for safe blasting. At Sungun Copper Mine (SCM), some unstable rock slopes facing to valuable industrial facilities are both expose to high intensity daily blasting vibrations, threatening their safty. So, controlling PPV by developing accurate predictors is essential. Hence, this study proposes improved strategies for prediction of PPV by maximum charge per delay and distance using the concept of Intelligent Committee Machine (ICM). Besides the Empirical Predictors (EPs) and two Artificial Intelligence (AI) models of ANFIS and ANN, four different ICMs models including Simple and Weighted Averaging ICM (SAICM and WAICM) and First and Second order Polynomial ICM (FPICM and SPICM) in conjunction with genetic algorithm, proposed for the prediction of PPV. Performance of predictors was studied considering R2, RSME and VAF indices. Results indicate that ICM methods have superiority over EPs, ANN and ANFIS, and among the ICM models while SAICM, WAICM and FPICM performing near to each other SPICM overrides all the models. R2 and RSME of the training and testing data for SPICM are 0.8571, 0.8352 and 11.0454, 12.3074, respectively. Finally, ICMs provides more accurate and reliable models rather than individual AIs.

Thermistors are very commonly used for narrow temperature-range high-resolution applications, such as in medicine, calorimetry, and near ambient temperature measurements. In particular, Negative Temperature Coefficient (NTC) thermistor is very inexpensive and highly sensitive, whose sensing temperature range and sensitivity are highly limited due to the intrinsic nonlinearity and self-heating properties of NTC thermistor at high operation currents. In this research, a new structure is proposed based on adaptive neuro-fuzzy system for the modeling of sensor nonlinearity. Apart from taking self-heating phenomenon of NTC thermistor sensor, the proposed structure also measures temperature directly, without any linearizing circuitry. Neuro-fuzzy network is trained and tested through data produced in the Laboratory environment. Examination of the proposed method on test data achieved a mean squared error of 0.0195, which is considered as a significant accomplishment.

This study investigates the potential of an intelligence model namely, Adaptive Neuro-Fuzzy Inference System (ANFIS) in prediction of the Zn metal ions adsorption in comparision with two well known empirical models included Thomas and Yoon methods. For this purpose, an organic-inorganic core/shell structure, γ-Fe2O3/polyrhodanine nanocomposite with γ-Fe2O3 nanoparticle as core with average diameter of 15 nm and polyrhodanine as shell with thickness of 3 nm, was synthesized via chemical oxidation polymerization. The properties of adsorbent were characterized with transmission electron microscope (TEM) and Fourier transform infrared (FT-IR) spectroscopy. Sixty seven experimental data sets including the treatment time (t), the initial concentration of Zn (Co), column height (h) and flow rate (Q) were used as input data to predict the ratios of effluent-to-influent concentrations of Zn (Ct/C0). The results showed that ANFIS model with the R coefficient of 0.99 can predict Ct/C0 more accurately than empirical models. Also it was found that the result of the Thomas and Yoon methods with R coefficient of 0.828 and 0.829, respectively were so close to each other. Finally, performance of our ANFIS model was compare to Thomas and Yoon methods in two different conditions, i.e. variable initial influent concentration and variable column height. High performance of ANFIS model was proved by the comparitive results.