Performance Of Artificial Neural Network Model Research Articles

Presentation of Machine Learning Approaches for Predicting the Severity of Accidents to Propose the Safety Solutions on Rural Roads

The aim of the current research was to develop models to predict the severity of accidents on rural roads in Tehran province, Iran. In this regard, using accident data from 2017 to 2020, the machine learning algorithms, including multiple logistic regression, multilayer perceptron neural network (MLPNN), and radial basis function neural network (RBFNN) models, as well as statistical methods, including Kolmogorov–Smirnov test, Friedman test, and factor analysis, were implemented to determine the contributory factors in the severity of accidents. Thus, nine variables affecting the severity of accidents were considered in modeling, and then the effect of each variable was calculated. By comparing the results of artificial neural network (ANN) models and the Friedman test, it was indicated that the human factor had a remarkable effect on accident severity. In addition, both machine learning and statistical methods can be served as guidance for safety authorities to provide safety solutions, thereby leading to reducing accidents. Finally, the performances of ANN models were analyzed by other mathematical models built by MATLAB programming.

Development and comparison of artificial intelligence models for estimating daily reference evapotranspiration from limited input variables

The accurate and efficient computational modelling framework for computation of reference evapotranspiration (ET0) is the need of time and is highly essential for several agricultural, hydrological and hydro-climatological studies and applications. ET0 has its own role especially for the water resource system management including irrigation water allocation and management, utilization and demand assessments of water resources. Considering the potential and wide applicability of Artificial Intelligence (AI) models in input-output mapping, these were examined in this research for modelling ET0 utilising a variety of meteorological data, including maximum temperature (Tmax) and minimum temperature (Tmin), relative humidity (RH), wind speed (WS), and bright sunshine hours (BSS). The different AI models investigated in this study include Artificial Neural Network (ANN), Extreme Learning Machine (ELM), M5 Tree and Multiple Linear Regression (MLR). The implemented daily dataset was collected from the Sardarkrushinagar station which is located at the North Gujarat region of India. For performance verification of the developed AI models, the FAO-56 Penman-Monteith equation was employed as a benchmark. The developed model's accuracy was assessed using multiple quantitative metrics, including R2, NSE, RMSE, Pdv and MAE. The performance of ANN models was found better compared to the ELM, M5 Tree, and MLR models in terms of different performance indicators. The model's outcomes were highly practical and reliable for the investigated case studies. Another important finding in this research was that the selection of the appropriate input variables in models allows not only error minimization but also enhances the co-relationship amongst the dependant as well as independent variables.

Artificial neural network analysis of the Nusselt number and friction factor of hydrocarbon fuel under supercritical pressure

This paper presents the Nusselt number and friction factor model for hydrocarbon fuel under supercritical pressure in horizontal circular tubes using an artificial neural network (ANN) analysis on the basis of the back propagation algorithm. The derivation of the proposed model relies on a large number of experimental data obtained from the tests performed with the platform of supercritical flow and heat transfer. Different topology structures, training algorithms and transfer functions are employed in model optimization. The performance of the optimal ANN model is evaluated with the mean relative error, the determination coefficient, the number of iterations and the convergence time. It is demonstrated that the model has high prediction accuracy when the tansig transfer function, the Levenberg-Marquardt training algorithm and the three-layer topology of 4-9-1 are selected. In addition, the accuracy of the ANN model is observed to be the highest compared with other classic empirical correlations. Mean relative error values of 4.4% and 3.4% have been achieved for modeling of the Nusselt number and friction factor respectively over the whole experimental data set. The ANN model established in this paper is shown to have an excellent performance in learning ability and generalization for characterizing the flow and heat transfer law of hydrocarbon fuel, which can provide an alternative approach for the future study of supercritical fluid characteristics and the associated engineering applications.

Estimating Quantity of Date Yield Using Soil Properties by Regression and Artificial Neural Network

ABSTRACT The objective of this study was to estimate the date production according to soil characteristics by two methods of multivariate regression and artificial neural networks. First, 90 groves with a wide range of soil characteristics but with almost the same climatic conditions and management were selected in five date-producing provinces of Iran. Then, a control profile was dug in each soil and the taken samples were sent to the laboratory for the designated physico-chemical analyses. Modeling was performed to estimate the quantity of production using soil variables with the highest correlation with yield. Obtained results revealed that some variables namely soil salinity, exchangeable sodium percentage, gypsum, lime, gravel, available potassium and phosphorus have highest correlation with the date yield. Comparison of two applied models indicated that both models are able to effectively estimate the date yield. The value of evaluation criteria consists of normalized root mean square error (NRMSE), coefficient of determination (R2) and coefficient of residual mass (CRM) for the artificial neural network model performance was better than the regression model. But, the average value of the predicted yield in the regression was more similar to the actual yield. Furthermore, breaking the data into different regions and creating multivariate regression with one or two important soil variables showed that the regression method can predict the quantity of date yield with acceptable accuracy. Consequently, due to the simplicity of implementation and application of multivariate regression model compared to artificial neural network and its acceptable accuracy, this model can be preferred.

Optimized Artificial Neural network models to time series

Artificial Neural networks (ANN) are powerful and effective tools in time-series applications. The first aim of this paper is to diagnose better and more efficient ANN models (Back Propagation, Radial Basis Function Neural networks (RBF), and Recurrent neural networks) in solving the linear and nonlinear time-series behavior. The second aim is dealing with finding accurate estimators as the convergence sometimes is stack in the local minima. It is one of the problems that can bias the test of the robustness of the ANN in time series forecasting. To determine the best or the optimal ANN models, forecast Skill (SS) employed to measure the efficiency of the performance of ANN models. The mean square error and the absolute mean square error were also used to measure the accuracy of the estimation for methods used. The important result obtained in this paper is that the optimal neural network was the Backpropagation (BP) and Recurrent neural networks (RNN) to solve time series, whether linear, semilinear, or non-linear. Besides, the result proved that the inefficiency and inaccuracy (failure) of RBF in solving nonlinear time series. However, RBF shows good efficiency in the case of linear or semi-linear time series only. It overcomes the problem of local minimum. The results showed improvements in the modern methods for time series forecasting.

Prediction of thirty-day morbidity and mortality after duodenal switch using an artificial neural network.

Understanding factors that increase risk of both mortality and specific measures of morbidity after duodenal switch (DS) is important in deciding to offer this weight loss operation. Artificial neural networks (ANN) are computational deep learning approaches that model complex interactions among input factors to optimally predict an outcome. Here, a comprehensive national database is examined for patient factors associated with poor outcomes, while comparing the performance of multivariate logistic regression and ANN models in predicting these outcomes. 2907 DS patients from the 2019 Metabolic and Bariatric Surgery Accreditation and Quality Improvement Program database were assessed for patient factors associated with the previously validated composite endpoint of 30-day postoperative reintervention, reoperation, readmission, or mortality using bivariate analysis. Variables associated (P ≤ 0.05) with the endpoint were imputed in a multivariate logistic regression model and a three-node ANN with 20% holdback for validation. Goodness-of-fit was assessed using area under receiver operating curves (AUROC). There were 229 DS patients with the composite endpoint (7.9%), and 12 mortalities (0.4%). Associated patient factors on bivariate analysis included advanced age, non-white race, cardiac history, hypertension requiring 3 + medications (HTN), previous foregut/obesity surgery, obstructive sleep apnea (OSA), and higher creatinine (P ≤ 0.05). Upon multivariate analysis, independently associated factors were non-white race (odds ratio 1.40; P = 0.075), HTN (1.55; P = 0.038), previous foregut/bariatric surgery (1.43; P = 0.041), and OSA (1.46; P = 0.018). The nominal logistic regression multivariate analysis (n = 2330; R2 = 0.02, P < 0.001) and ANN (R2 = 0.06; n = 1863 [training set], n = 467 [validation]) models generated AUROCs of 0.619, 0.656 (training set) and 0.685 (validation set), respectively. Readily obtainable patient factors were identified that confer increased risk of the 30-day composite endpoint after DS. Moreover, use of an ANN to model these factors may optimize prediction of this outcome. This information provides useful guidance to bariatricians and surgical candidates alike.

Prediction of the drying shrinkage of alkali-activated materials using artificial neural networks

Alkali-activated materials (AAMs) are qualitatively and quantitatively evaluated with an emphasis on the ultimate drying shrinkage. We systematically evaluated AAMs based on the mix design and curing conditions, utilizing a total of 452 AAM mixtures extracted from 44 papers. Finally, a predictive model for the ultimate drying shrinkage of AAMs was constructed using an artificial neural network (ANN) with high accuracy, in which the reactivity of binder, geopolymer paste volume, liquid-to-binder ratio, alkali activator modulus, aggregate volumetric ratio, curing temperature, relative humidity and specimen size were set as inputs. This model shows great generality by compiling various AAM mixtures and is easy-handling without preparation of samples for acquiring specific properties. Moreover, the efficiency of three commonly used models for predicting the drying shrinkage—the Bažant-Baweja model, Gardner and Lockman model, and multi-linear regression model—were evaluated and compared to the proposed ANN model, revealing a better prediction performance of ANN model. This study will advance the understanding of the drying shrinkage behaviors of AAMs and provide practical guidelines for designing AAM mixtures with high durability.

Effect of input-target datasets on sediment transport modelling in alluvial rivers using artificial neural network

ABSTRACT Numerical simulation of sediment transport in alluvial rivers often becomes untenable due to the absence of requisite data. Over the years, artificial neural network (ANN) models have been successfully employed for tackling such scenarios. However, the effect of input and target datasets on the simulation of sediment transport using ANN has not yet been adequately addressed in the literature so far. To study the effect of input-target dataset on the performance of ANN models, the present study employs seven input-target datasets for the development of several ANN models and assessment of their performance. The limited study carried out reveals that there are multiple ANN models – having different combinations of training algorithms, transfer functions and input/target datasets – that can reliably be employed to estimate bed elevation changes in alluvial rivers.

Modelling of Reference Evapotranspiration using Artificial Neural Network in Semi-arid Region of North Gujarat

In order to efficiently manage the limited water resources in semi-arid regions, accurate quantification of irrigation water requirement of irrigated areas is deemed necessary. Since the irrigation water requirements of crops depend on reference evapotranspiration (ET0 ), assessment of ET0 is a key task in development of water resource management plans. This study examined effectiveness of the use of artificial neural networks (ANN) for the estimation of ET0 using incomplete meteorological variables. These ANN models use daily meteorological data (maximum and minimum temperature, relative humidity, wind speed, and bright sunshine hours) as inputs, and provide ET0 as outputs. Considering individual and possible combinations of these 5 input variables in combination with different number of hidden neurons (1 to 20), 31 model architectures were evaluated for their accuracy in estimating ET0 . The results showed that the proper choice of neural network architecture allowed not only error minimization, but also strengthened the relationship between the dependent variable and the independent variables. The ANN architecture 5-11-1, having all five parameters as input, showed highest accuracy with R2 , NSE, RMSE, MAE as 0.98, 98.16%, 0.24 mm.day-1, 0.17 mm.day-1 during training; and 0.98, 98.11%, 0.27 mm.day-1, 0.19 mm.day-1 during the testing period, respectively. Comparative performance of ANN models showed choice of input meteorological variables and model architecture are major determinants of ANN model performance and that the properly trained and tested ANN model can be used in estimation of ET0 in the semi-arid regions of India.

An improved approach towards more robust deep learning models for chemical kinetics

The application of artificial neural network (ANN) models in replacement of numerical integration solvers for ordinary differential equations (ODE) has attracted increasing interest in recent years owing to its high computational speed potential and approximation capabilities. However, regarding its application for the calculation of chemical kinetics in combustion simulations, there are challenges over its suitability in wide-range reaction conditions, error propagation over excessive time steps, and its performance in cases with high nonlinearity (especially at elevated pressures). In this study, a deep learning model was successfully developed as a fast alternative to the numerical integration of the detailed mechanism of hydrogen for time-resolved ignition simulations at pressures up to 45 atm, which is featured by the abruptness of ignition and long ignition delays. More importantly, some methodological improvements are proposed and discussed in an effort to build more robust models and to extend the applicable range of the deep learning model without sacrificing model compactness. These techniques include the employment of continuously differentiable activation functions, rate-distinguishing classification, and perturbed simulation for data generation. This study shows that these techniques can unlock further potentials of ANN in making reliable predictions in practical applications. The fair comparison between these methods and conventional ones in this study also enables a deeper understanding of the effect of data distribution and network architecture on the performance of ANN models in solving stiff ODEs.

DeepDRG: Performance of Artificial Intelligence Model for Real-Time Prediction of Diagnosis-Related Groups.

Nowadays, the use of diagnosis-related groups (DRGs) has been increased to claim reimbursement for inpatient care. The overall benefits of using DRGs depend upon the accuracy of clinical coding to obtain reasonable reimbursement. However, the selection of appropriate codes is always challenging and requires professional expertise. The rate of incorrect DRGs is always high due to the heavy workload, poor quality of documentation, and lack of computer assistance. We therefore developed deep learning (DL) models to predict the primary diagnosis for appropriate reimbursement and improving hospital performance. A dataset consisting of 81,486 patients with 128,105 episodes was used for model training and testing. Patients’ age, sex, drugs, diseases, laboratory tests, procedures, and operation history were used as inputs to our multiclass prediction model. Gated recurrent unit (GRU) and artificial neural network (ANN) models were developed to predict 200 primary diagnoses. The performance of the DL models was measured by the area under the receiver operating curve, precision, recall, and F1 score. Of the two DL models, the GRU method, had the best performance in predicting the primary diagnosis (AUC: 0.99, precision: 83.2%, and recall: 66.0%). However, the performance of ANN model for DRGs prediction achieved AUC of 0.99 with a precision of 0.82 and recall of 0.57. The findings of our study show that DL algorithms, especially GRU, can be used to develop DRGs prediction models for identifying primary diagnosis accurately. DeepDRGs would help to claim appropriate financial incentives, enable proper utilization of medical resources, and improve hospital performance.

Modeling yield and backscatter using satellite derived biophysical variables of rice crop based on Artificial Neural Networks

In this study, ‘observed rice yield (ton acre-1)’ and ‘remotely sensed backscatter’are modelled using artificial neural network (ANN) and multiple linear regression (MLR) methods for East and West Godavari districts of Andhra Pradesh in India. The biophysical variables viz. backscatter (bs), normalized difference vegetation index (NDVI), Chlorophyll (chfl), fraction of absorbed photosynthetically active radiation (FAPAR), leaf area index (LAI), canopy water content (CWC), and fraction of vegetation cover (Fcover) were derived from Scatterometer Satellite-1 (SCATSAT-1), Moderate Imaging Spectrometer (MODIS) and Sentinel-2 satellite data.Inputs selected are bs, NDVI, chfl, FAPAR, LAI, CWC, and Fcover for rice yield model, whereas NDVI, chfl, FAPAR, LAI, CWC, and Fcover are inputs for backscatter models. The performance of ANN and MLR models was evaluated using three indices such as root mean squared error (RMSE), mean absolute error (MAE), and coefficient of determination (R2). The results concluded that the ANN models achieved R2 of 0.908 and 0.884 which are 42.73% and 28.85% higher than that of the MLR method for rice yield and backscatter, respectively.

Prediction of spark ignition engine performance responses fueled with fusel oil/gasoline blends by artificial neural network

In the present study, the performance parameters of a single-cylinder, air-cooled spark ignition (SI) engine using fusel oil-gasoline fuel blends were predicted by artificial neural network (ANN). The SI engine was operated with gasoline/fusel oil (10% and 20%) blends at different engine load (1000, 2000, 3000, 4000, 5000, 6000, 7000 and 8000 Watt) and compression ratios (8.00, 9.12 and 10.07) to obtain data essential to create the ANN model. In the constructed ANN model, brake thermal efficiency (BTE) and brake specific fuel consumption (BSFC) are chosen as output parameters, while engine load, compression ratio (CR) and fusel oil ratio are chosen as input factors. 75% of the test results were employed to train the ANN. The performance of ANN model was determined by comparing it with the data produced from the part not used for training. According to the found data, ANN model estimated engine performance parameters such as BTE and BSFC by an overall regression coefficient (R) at 0.99384. Simultaneously, mean absolute percentage error (MAPE) were found as 5.027% and 7.847% for BTE and BSFC, respectively. When ANN results and experimental results are compared for BTE and BSFC responses, it is determined that ANN results are close to experimental results with an error rate of less than 5%.

PSVI-16 Comparison of univariate and multivariate machine learning applications for genomic predictions of growth and carcass traits from Brangus heifers

Abstract Data for growth and carcass traits were obtained from 738 Brangus heifers (3/8 Brahman-Bos indicus × 5/8 Angus-Bos taurus) registered with International Brangus Breeders Association. Phenotypes included body weights (BW, WW, and YW) collected at birth, ~205 and 365 d of age for growth traits and yearling ultrasound assessment of longissimus muscle area (LMA), percent intramuscular fat (IMF), and depth of rib fat (FAT) for carcass traits and were used to compare univariate and multivariate artificial neural networks (ANN) models with the learning algorithms (Levenberg–Marquardt (LM), Bayesian Regularization (BR) and Scaled Conjugate Gradient (SCG)) and transfer functions (tangent sigmoid and linear) using 1 to 20 neurons models based on the input from G genomic relationship matrix. Pearson correlation coefficients for evaluating model performances in testing datasets indicated that univariate ANN models resulted in better genomic prediction than the multivariate ANN models regardless of the learning algorithms and transfer functions for carcass traits. However, there was no clear difference between univariate and multivariate ANN models about genomic prediction for growth traits. In multivariate ANN models, the prediction performance of the combination of learning algorithms and transfer functions changed for growth and carcass traits and there were no superior multivariate ANN models with BR, LM and SCG learning algorithms and transfer functions. However, the correlation coefficients from univariate ANN-BR model indicated better genomic prediction than univariate ANN-LM and ANN-SCG models. The application of different transfer functions did not make any significant difference on the genomic prediction performance of ANN models with different learning algorithms. Results of this study suggest the use of univariate ANN models with BR learning algorithm for genomic prediction of growth and carcass traits.

Prediction of storage time in different seafood based on color values with artificial neural network modeling.

The determination of storage time in seafood could be performed by microbiological, chemical and sensory analysis. Among these mentioned methods color changes are one part of sensory analysis and are prior acceptance criteria from the point of consumers' view. In this study, a feedforward artificial neural network (ANN) model was developed to predict the storage time of seafood based on L*, a* and b* values. A total of 205 data set were compiled from the literature that represents the color changes of different seafood products to train and test the ANN model. Another set of data (n = 45) were used for the validation of developed ANN model. A multi-layer perceptron (MLP) was applied for the determination of agreements between input and output data. The most accurate topology were determined in accordance with the changes in the values of correlation coefficients (R2) and mean square errors (MSE) and found to be 30 neurons in the layer (R2 = 0.81 and MSE = 0.2). The performance of ANN model was evaluated based on 6 criteria such as Mean Absolute Deviation (MAD), Mean Square Errors (MSE), Residual Mean Square Errors (RMSE), Correlation Coefficient (R2), Mean Absolute Error (MAE) and F-test statistics and found to be 0.2, 0.05, 0.002, 0.8, 0.71 and 1.06, respectively. Moreover, predicted and observed storage time values were fitted and regression coefficient was found to be 0.85. In accordance with the results of this study, the proposed ANN model is accurate, reliable, and proper for the estimation of storage time in seafood products.

Experimental investigation and artificial neural network modeling of performance and emission of a CI engine using orange peel oil- diesel blends

ABSTRACT The exponential decay of crude oil and environmental degradation forced researchers to investigate alternative sustainable fuels. One of the promising alternatives to petroleum oil is orange peel oil. Orange peel oil has characteristics that are similar to diesel, and it is also simple to blend. In this paper, performance and emission parameters are investigated, and using an artificial neural network (ANN), its prediction and validation have been made. The blend shows the best results for 30% orange peel oil by vol. and 70% diesel. At peak load, the brake thermal efficiency (BTE) of this blend is 17.5% higher than diesel, and the brake-specific energy consumption (BSEC) is 20% lower. These experimental results were used to perform prediction and validation. The prediction of data is made employing the Quasi- newton method algorithm, and it is found that it best fits the linear regression analysis. The value of R2 for BTE and BSEC is 0.994 and 0.986, respectively. The suggested ANN model performance and accuracy were totally satisfactory. There is a noticeable reduction in the emission parameters.

Designing artificial neural network of nanoparticle diameter and solid–fluid interfacial layer on single‐walled carbon nanotubes/ethylene glycol nanofluid flow on thin slendering needles

AbstractIn this study, an artificial neural network (ANN) has been developed to predict the boundary layer flow of a single‐walled carbon nanotubes nanofluid toward three different nonlinear thin isothermal needles of paraboloid, cone, and cylinder shapes with convective boundary conditions. Different effects of particle diameter and solid–fluid interface coating have been taken into account in the thermal conductivity model of nanofluid in which ethylene glycol has been used as the base fluid. Single and dual phase approach is used to establish the management model under the phenomenon of zero heat and mass flux. A dataset has been developed for different scenarios of the fluid model by changing the relevant parameters with the Runge–Kutta based shooting technique. Two different ANN models have been developed to predict Nusselt number and skin friction coefficient (SFC) values. The values obtained from ANN models have been compared with the numerical data, which are the target values. In addition, mean square error and R values have also been examined in order to analyze the prediction performance of ANN models more comprehensively. The calculated R values for Nusselt number and SFC were obtained as 0.9999. The results obtained showed that ANN can predict Nusselt number and SFC values with high accuracy.

Prediction of heat transfer distribution induced by the variation in vertical location of circular cylinder on Rayleigh-Bénard convection using artificial neural network

The present study investigates the flow and thermal fields of Rayleigh-Bénard convection (RBC) in a rectangular channel with an internal circular cylinder. The parameters considered are Rayleigh number (104≤Ra≤106), Prandtl number (Pr = 0.7), and irreversibility distribution ratio (φ = 1). The vertical distance (δ) in the range of -0.2 ≤ δ ≤ 0.2 is the major simulation parameter in present study. The results are analyzed based on the iso-surface of temperature, vortical structure with orthogonal enstrophy distribution, and entropy generations. Additionally, Nusselt number (Nu) and Bejan number (Be) are obtained to analyze the heat transfer characteristics and irreversibility, respectively. The Rayleigh number and the vertical distance significantly influence the flow and thermal characteristics within the channel. Besides, an artificial neural network (ANN) model is used to predict the distribution of local Nusselt number. The performance of present ANN model is evaluated by comparing the tendency and quantitative values with the direct numerical simulation (DNS) results. The results show that the ANN model used in this study can precisely predict the correlation between the input parameters and output parameter with lesser computational time and cost compared to the DNS.

Estimation of unsteady hydromagnetic Williamson fluid flow in a radiative surface through numerical and artificial neural network modeling

In current investigation, a novel implementation of intelligent numerical computing solver based on multi-layer perceptron (MLP) feed-forward back-propagation artificial neural networks (ANN) with the Levenberg–Marquard algorithm is provided to interpret heat generation/absorption and radiation phenomenon in unsteady electrically conducting Williamson liquid flow along porous stretching surface. Heat phenomenon is investigated by taking convective boundary condition along with both velocity and thermal slip phenomena. The original nonlinear coupled PDEs representing the fluidic model are transformed to an analogous nonlinear ODEs system via incorporating appropriate transformations. A data set for proposed MLP-ANN is generated for various scenarios of fluidic model by variation of involved pertinent parameters via Galerkin weighted residual method (GWRM). In order to predict the (MLP) values, a multi-layer perceptron (MLP) artificial neural network (ANN) has been developed. There are 10 neurons in hidden layer of feed forward (FF) back propagation (BP) network model. The predictive performance of ANN model has been analyzed by comparing the results obtained from the ANN model using Levenberg-Marquard algorithm as the training algorithm with the target values. When the obtained Mean Square Error (MSE), Coefficient of Determination (R) and error rate values have been analyzed, it has been concluded that the ANN model can predict SFC and NN values with high accuracy. According to the findings of current analysis, ANN approach is accurate, effective and conveniently applicable for simulating the slip flow of Williamson fluid towards the stretching plate with heat generation/absorption. The obtained results showed that ANNs are an ideal tool that can be used to predict Skin Friction Coefficients and Nusselt Number values.

Prediction of Discharge Using Artificial Neural Network and IHACRES Models Due to Climate Change

Understanding of climate change and its impacts on river discharge has affected the quality and quantity of water and also supplying water requirements for drinking, agriculture and industry. Therefore, prediction of precipitation and temperature by climate models as well as simulation and optimization of their runoff with suitable models are very important. In this study, four climate models of the Fifth Coupled Model Inter comparison Project (CMIP5) and RCP8.5 scenario were used to forecast future precipitation and temperature for the next two periods including 2020-2052 and 2053-2085. Mean Observed Temperature-Precipitation (MOTP) method was used to reduce the uncertainty of climate models and the change factor method was used to downscale the climate data. Then, the Lumped-conceptual Identification of unit Hydrographs and Component flows from Rainfall, Evaporation and Stream flow data (IHACRES) model and multi-layer Artificial Neural Network (ANN) model were employed to estimate the effects of these parameters on the Khorramrood River runoff. The neural network model is written and implemented using Scikit-Learn library and the Python programming language. The comparison of performance of ANN models with different input variables like monthly precipitation, monthly precipitation of previous months, monthly discharge, monthly discharge of previous months, monthly temperature was made to find the best and most efficient network structure. The results showed that the precipitation in Khorramrood River basin based on the weighted combination model decreased by 8.18 % and 9.75 % in the first and the second periods, respectively, while the temperature increased by 1.85 and 4.22 °C, respectively. The discharge parameter in the calibration and validation period in the IHACRES model based on criteria to evaluate the parameters of Root Mean Square Error (RMSE), Mean Absolute Error (MAE), The Coefficient of Determination (R), and the Nash-Sutcliffe Efficiency (NSE) performed better than the artificial neural network model. However, due to the small differences of these changes, the predictions were performed for both periods and using both models and the results indicated that future discharge in the IHACRES model decreased by 12.72 % during the first period and by 20.3 % in the second period, while the model of artificial neural network showed decrease rates of 2.12 % and 6.97 %, respectively.