Historical researches concerning unit hydrograph (UH) demonstrate significant variation and presence of noise due to error in measurements and the assumption of linearity in rainfall–runoff relation. In this study, first, the UH is calculated for 16 storm events based on The least square method in a 449 hectare basin located in the vicinity of Riesel city, USA. In order to recognize noise frequencies, spectrums of the UHs are calculated for each UH by Fourier transformation. The results obtained from the spectral analysis are used for the noise reduction of the UH via a normal - weighted moving average method. From 16 smoothed UHs, one UH is selected as the representative UH of the basin by crossvalidation method. The results showed that the methods of the weighted moving average and cross– validation are suitable for obtaining a non – oscillatory and unique UH.

Groundwater is one of the major sources of exploitation in arid and semi-arid regions. Thus for protecting groundwater quality, spatial and temporal distribution of data are important. The objective of this research was to evaluate interpolation methods for predicting spatial distribution of some groundwater quality such as TDS, Na+, EC, SAR, Cl- and SO4-2. Data related to 65 wells in Rafsanjan plain were used and IDW, kriging and cokriging were studied. After normalization of data, variograme was drawn. For selecting suitable model for fitness on experimental variograme, less RSS value was used. By using crossvalidation and RMSE, the best method for interpolation was then selected. Results showed that for interpolation of groundwater quality, kriging and cokriging methods are superior to IDW method. At last, the maps of groundwater quality were prepared, using the best interpolation method in GIS environment.

This research was conducted to investigate the spatial structure to estimate the crown cover and density of a coppice oak forest in west of Iran (Loristan province) using Kriging and IDW (Inverse Distance Weighting) interpolation methods. Field sampling was performed based on a 100m×100m systematic grid using 1500 m2 circular samples of. Totally, 54 sample plots were measured at 54 ha. Experimental variograms for forest stem density and crown cover were calculated and plotted using the geo-referenced inventory plots. The calculated variograms of stem density and crown cover showed medium spatial autocorrelation fitted by spherical models. Estimations were made by ordinary block (38m×38m) kriging and IDW (power=2). Crossvalidation results showed that all estimations are unbiased. Therefore, Kriging and IDW are able to accurately estimate and map the crown cover and density of this kind of coppice forests.

Vegetation is an important component of global ecosystem and the knowledge about its cover and fraction are essential in understanding of land-atmosphere interactions and their effects on environmental issues. The main objective of this study was to estimate the vegetation fraction (Fv) of an arid area in central part of Iran (Sheitoor- Yazd) by using satellite images and artificial neural networks (ANN). To do so, the percentage of vegetation fraction for 52 randomly selected plots (50 meters by 50 meters), were measured on the field in July 2009. Next, an ALOS (AVNIR) image collected on 18 July 2009 and multilayer perceptron network were used to estimate the percentage of vegetation cover. Two types of transfer function, 12 training functions and six different combinations of spectral bands of satellite image as input were used to select the optimal network. Furthermore, the number of hidden neurons varied from one to six. Field measurements were used as target values to the network. To evaluate the effect of randomly selected training and test data, 30 and 35 out of 52 observed plots were considered as training data sets, and 22 and 17 plots as test data sets, respectively. Then, using linear regression models between the measured field data and estimated values, coefficients of determinations and RMSEs were calculated. Moreover in order to validate the results and remove possible errors due to random selection, cross validation algorithm was used. Results demonstrate that ANN can be used for accurate estimation of the percentage of vegetation cover in arid areas (R2>0.74 and RMSE <2%).

In the present study, we used 27 precipitation average monthly data from synoptic, climatologic, rain-guage and evaporative stations located in Zayandeh-Rud river basin for the period of 1970-2014. Before interpolating, the missing data in the time series of each station was reconstructed by the normal ratio method. Also, for the data quality control, the Dickey-Fuller and Shapiro-Wilk tests were used to check the data stationarity and normality. Then, these data were interpolated by six interpolation methods including Inverse Distance Weighting, Natural Neighbor, Tension Spline, Regularized Spline, Ordinary Kriging and Universal Kriging; then each method was evaluated using the crossvalidation technique with MAE, MBE and RMSE indices. The results showed that among the spatial interpolation methods, Natural Neighbor method with MAE of 0. 24 had the best performance for interpolating precipitation among all of the methods. Also, among Ordinary Kriging, Universal Kriging, Spline and Inverse Distance Weighting methods, respectively, Exponential Kriging with MAE 0. 54, Quadratic Drift Kriging with MAE of 0. 5, Tension Spline with the MAE of 0. 54 and Inverse Distance Weighting with the power of 4 with MAE of 0. 57 had the least error compared to other IDW methods.

Background: Nucleic acid-binding proteins play major roles in different biological processes, such as transcription, splicing and translation. Therefore, the nucleic acidbinding function prediction of proteins is a step toward full functional annotation of proteins. The aim of our research was the improvement of nucleic-acid binding function prediction. Methods: In the current study, nine machine-learning algorithms were used to predict RNA-and DNA-binding proteins and also to discriminate between RNA-binding proteins and DNA-binding proteins. The electrostatic features were utilized for prediction of each function in corresponding adapted protein datasets. The leave-one-out crossvalidation process was used to measure the performance of employed classifiers. Results: Radial basis function classifier gave the best results in predicting RNA-and DNA-binding proteins in comparison with other classifiers applied. In discriminating between RNA-and DNA-binding proteins, multilayer perceptron classifier was the best one. Conclusion: Our findings show that the prediction of nucleic acid-binding function based on these simple electrostatic features can be improved by applied classifiers. Moreover, a reasonable progress to distinguish between RNA-and DNA-binding proteins has been achieved.

Background: A timely diagnosis of Alzheimer’ s disease (AD) is crucial to obtain more practical treatments. In this article, a novel approach using Auto‑ Encoder Neural Networks (AENN) for early detection of AD was proposed. Method: The proposed method mainly deals with the classification of multimodal data and the imputation of missing data. The data under study involve the MiniMental State Examination, magnetic resonance imaging, positron emission tomography, cerebrospinal fluid data, and personal information. Natural logarithm was used for normalizing the data. The Auto‑ Encoder Neural Networks was used for imputing missing data. Principal component analysis algorithm was used for reducing dimensionality of data. Support Vector Machine (SVM) was used as classifier. The proposed method was evaluated using Alzheimer’ s Disease Neuroimaging Initiative (ADNI) database. Then, 10fold crossvalidation was used to audit the detection accuracy of the method. Results: The effectiveness of the proposed approach was studied under several scenarios considering 705 cases of ADNI database. In three binary classification problems, that is AD vs. normal controls (NCs), mild cognitive impairment (MCI) vs. NC, and MCI vs. AD, we obtained the accuracies of 95. 57%, 83. 01%, and 78. 67%, respectively. Conclusion: Experimental results revealed that the proposed method significantly outperformed most of the stateoftheart methods.

Background: Pulmonary movements during radiation therapy can cause damage to healthy tissues. It is necessary to adapt treatment planning based on tumor motion to avoid damage to healthy tissues. A range of approaches has been proposed to monitor the issue. A treatment planning based on fourdimensional computed tomography (4D CT) images can be addressed as one of the most achievable options. Although several methods proposed to predict pulmonary movements based on mathematical algorithms, the use of deep artificial neural networks has recently been considered. Methods: In the current study, convolutional long shortterm memory networks are applied to predict and generate images throughout the breathing cycle. A total of 3295 CT images of six patients in three different views was considered as reference images. The proposed method was evaluated in six experiments based on a leaveonepatientout method similar to crossvalidation. Results: The weighted average results of the experiments in terms of the rootmeansquared error and structural similarity index measure are 9 × 10^− 3 and 0. 943, respectively. Conclusion: Utilizing the proposed method, because of its generative nature, which results in the generation of CT images during the breathing cycle, improves the radiotherapy treatment planning in the lack of access to 4D CT images.

Many hydrological models require high resolution rainfall data. The aim of this paper was to interpolate annual and monthly rainfall in Golestan province from sparse point data. To do this the methods, which make use of secondary variables (e.g. a digital elevation model, DEM) for rainfall estimation were compared with those, which do not make use of such information in estimation. The methods applied included univariate interpolation algorithms such as inverse square distance and ordinary kiriging and multivariate geostatistical algorithms such as cokriging, kriging with an external drift and simple kriging with varying local means. The performance of each interpolator was assessed through examination of mapped estimates of rainfall and crossvalidation. It was concluded that cokriging provides the most accurate estimates of rainfall for May to October except June which was best estimated using kiriging with an external drift judging by the CROSS-VALIDATION estimation error summary statistics. For other periods ordinary kriging yielded more accurate rainfall predictions than other interpolators. The worst algorithm was inverse square distance that ignores both the elevation and rainfall records at neighboring stations. Relative nugget effect of semivariograms and correlation between rainfall and elevation affected the performance of different methods. For instance, ordinary kriging outperformed other technique when the correlation between rainfall and elevation was less than 0.5.