応用水文 = Applied hydrology

Research highlights? The use of multiple regression (MR), artificial neural network (ANN) and artificial neuro-fuzzy inference system (ANFIS) models, for the prediction of swell percent of soils, was described and compared. ? However the accuracies of ANN and ANFIS models may be evaluated relatively similar, it is shown that the constructed ANN models of RBF and MLP exhibit a high performance than ANFIS and multiple regression for predicting swell percent of clays. ? The performance comparison showed that the soft computing system is a good tool for minimizing the uncertainties in the soil engineering projects. ? The use of soft computing will also may provide new approaches and methodologies, and minimize the potential inconsistency of correlations. In the recent years, new techniques such as; artificial neural networks and fuzzy inference systems were employed for developing of the predictive models to estimate the needed parameters. Soft computing techniques are now being used as alternate statistical tool. Determination of swell potential of soil is difficult, expensive, time consuming and involves destructive tests. In this paper, use of MLP and RBF functions of ANN (artificial neural networks), ANFIS (adaptive neuro-fuzzy inference system) for prediction of S% (swell percent) of soil was described, and compared with the traditional statistical model of MR (multiple regression). However the accuracies of ANN and ANFIS models may be evaluated relatively similar. It was found that the constructed RBF exhibited a high performance than MLP, ANFIS and MR for predicting S%. The performance comparison showed that the soft computing system is a good tool for minimizing the uncertainties in the soil engineering projects. The use of soft computing will also may provide new approaches and methodologies, and minimize the potential inconsistency of correlations.

Multiple regression, ANN (RBF, MLP) and ANFIS models for prediction of swell potential of clayey soils

Modeling And Optimizing of Different Quality Characteristics In Electrical Discharge Drilling Of Titanium Alloy (Grade-5) Sheet

The Water Quality Index (WQI) is the most common indicator to characterize surface water quality. This study introduces a new ensemble machine learning model called Extra Tree Regression (ETR) for predicting monthly WQI values at the Lam Tsuen River in Hong Kong. The ETR model performance is compared with that of the classic standalone models, Support Vector Regression (SVR) and Decision Tree Regression (DTR). The monthly input water quality data including Biochemical Oxygen Demand (BOD), Chemical Oxygen Demand (COD), Dissolved Oxygen (DO), Electrical Conductivity (EC), Nitrate-Nitrogen ( NO 3 -N), Nitrite-Nitrogen ( NO 2 -N), Phosphate ( P O 4 3 - ), potential for Hydrogen (pH), Temperature (T) and Turbidity (TUR) are used for building the prediction models. Various input data combinations are investigated and assessed in terms of prediction performance, using numerical indices and graphical comparisons. The analysis shows that the ETR model generally produces more accurate WQI predictions for both training and testing phases. Although including all the ten input variables achieves the highest prediction performance ( R 2 t e s t = 0.98 , R M S E t e s t = 2.99 ), a combination of input parameters including only BOD, Turbidity and Phosphate concentration provides the second highest prediction accuracy ( R 2 t e s t = 0.97 , R M S E t e s t = 3.74 ). The uncertainty analysis relative to model structure and input parameters highlights a higher sensitivity of the prediction results to the former. In general, the ETR model represents an improvement on previous approaches for WQI prediction, in terms of prediction performance and reduction in the number of input parameters.

River water quality index prediction and uncertainty analysis: A comparative study of machine learning models

In this paper, Infiltration rate of the soil is investigated by using predictive models of Random forest regression and their performance were compared with Artificial neural network (ANN) and M5P model tree techniques. A dataset consists of 132 field measurements were used. Out of 132 observations randomly selected 88 observations were used for training, whereas remaining 44 were used for testing the model. Input variables consist of cumulative time (Tf), type of impurities (It), concentration of impurities (Ci), and moisture content (Wc) whereas the infiltration rate was considered as output. Correlation coefficient (CC), root mean square error (RMSE), mean absolute error (MAE), relative absolute error (RAE) and root relative square error (RRSE) were considered to compare the performance the both modelling approaches. The result of evolution suggests that Random forest regression approach works well than the other two models (ANN and M5P model tree). The estimated value of infiltration rate using Random forest regression lies within ±25% error lines. Sensitivity analysis suggests that cumulative time is an important parameter for predicting the infiltration rate of the soil.

Modelling of impact of water quality on infiltration rate of soil by random forest regression

This paper aims to predict the unsaturated hydraulic conductivity of soil using Adaptive Neuro- fuzzy inference system (ANFIS), Multi-Linear Regression (MLR), and artificial neural network (ANN). Laboratory experiments carried out on 46 samples of sand, rice husk ash and fly ash (FA) mixture. Out of 46 data-set for modeling of unsaturated hydraulic conductivity 32 random data used for training and remaining 14 to the test. The results suggest improved performance by Gaussian membership function than triangular and generalized bell-shaped membership-based ANFIS. MLR is better than ANN and Gaussian membership function-based ANFIS for unsaturated hydraulic conductivity.

Prediction of unsaturated hydraulic conductivity using adaptive neuro- fuzzy inference system (ANFIS)

Support vector regression based modeling of pier scour using field data

The aim of this paper to assesses the potential of machine learning approaches, i.e. multi-linear regression (MLR), support vector regression (SVR), Gaussian process (GP) regression of cumulative infiltration and compares their performances with three traditional models [Kostiakov model, US-Soil Conservation Service (SCS) model and Philip’s model]. Data set as many as 413 were obtained by conducting experiments in laboratory of NIT Kurukshetra. It is observed from the experiments that moisture content influences the cumulative infiltration of soil. Out of 413 data set 289 arbitrary selected were used for training the models, whereas remaining 124 were used for testing. Input data set consist of time, sand, rice husk ash, fly ash, suction head, bulk density and moisture content where as cumulative infiltration was considered as output. Two kernel function i.e. Pearson VII and radial based kernel function were used with both SVR and GP regression. The results after comparison suggests that the GP regression based approach works better than SVR, MLR, Kostiakov model, SCS model and Philip’s model approaches and it could be successfully used in prediction of cumulative infiltration data.

Modelling of infiltration of sandy soil using gaussian process regression

Infiltration models are very helpful in designing and evaluating surface irrigation systems. The main objective of the present work is estimation and inter-comparison of infiltration models which are used to evaluate the infiltration rates of National Institute of Technology (NIT)-campus in district of Kurukshetra, Haryana (India) and for this study, field infiltration tests were carried out at ten different locations comprising of 109 observations by use of double ring infiltrometer. The potential of three infiltration models (Kostiakov, Modified Kostiakov and US- Soil Conservation Service (SCS)) were evaluated by least–square fitting to observed infiltration data. Three statistical comparison criteria including maximum absolute error (MAE), Bias and root mean square error (RMSE) were used to determine the best performing infiltration models. In addition, a novel infiltration model was developed from field tests data using nonlinear regression modeling which suggests improved performance out of other three models. In case of nonexistence of observed infiltration data, this novel model can be used to artificially generate infiltration data for NIT campus.

Estimation and inter-comparison of infiltration models

This paper examines the capability of the support vector machine-based regression approach to predict the cumulative infiltration from sandy soil. A data-set consisting of 413 cumulative in...

N. K. Tiwari

Papers

Prediction of unsaturated hydraulic conductivity using adaptive neuro- fuzzy inference system (ANFIS)

Support vector regression based modeling of pier scour using field data

Modelling of infiltration of sandy soil using gaussian process regression

Estimation and inter-comparison of infiltration models

Support vector regression-based modeling of cumulative infiltration of sandy soil