learning algorithm

Precipitation is considered one of the most important components of hydrological cycle, and its effective and usable amount for plants is of great importance in the agricultural sector, especially rainfed cultivation. In this research, the effective precipitation (EP) in dry wheat fields of Khomein city was estimated by using RS and SEBAL on 28 available images from Landsat8 in the crop years 2014 to 2022. Penman-Monteith-Fao method was used to evaluate the accuracy of SEBAL. Then, a model of EP estimation was developed with ANN and meteorological data. For this purpose, the correlation between meteorological data and Growin Degree Days (GDD) with EP was investigated by Pearson's correlation method. the meteorological data of three stations from the closest synoptic stations to the study area were used and The meteorological data of the study area were interpolated using the Inverse Distance Weighting method (IDW). According to the results of the correlations, the average temperature parameter with a correlation of 0.92 and the GDD and the maximum relative humidity respectively with a correlation of 0.86 and -0.77 as effective variables in estimating EP. In the next step, the most effective parameters were used for modeling. the networks were trained under different scenarios, and the performance of the networks was evaluated using the RMSE and MBE error criteria. The results showed that by using the BR learning algorithm and having the variables of daily temperature and GDD, it is possible to predict the amount of EP for the target area with very good accuracy. The RMSE value of this model was 0.1899 mm and MBE was estimated as -0.0115 mm. By using the presented model, with simple meteorological variables, the actual evapotranspiration and finally the EP of the desired area can be determined with appropriate accuracy without the need to solve complex algorithms.

Appropriate selection of ancillary covariates have a specific important on digital soil mapping. Currently, use of machine learning algorithms for digital mapping and updating of conventional soil map has been developed in Iran. The current study has been done to compare the BRT and RF models for spatial prediction of subgroup and family classes with selection of axillary variables  using VIF approach in some part of Qazvin Plain. 61 pedons were sampled based on stratified random, digged, described and classified with consideration of laboratory analysis up to family level. The most appropriate variables were selected among 15 Geomorphometry and Remote Sensing Indices using Variance Inflation Factor (VIF). Soil landscape modeling was conducted with RF and BRT learning algorithm in RStudio software based on Randomforest and C5.0 packages at subgroup and family levels. The results showed that six indices including CHA, DEM, STH, SI DVI and NDVI were selected as input variables. Assessment indices such as the Overall Accuracy (OA) and Kappa were obtained for BRT (35, 26%) and RF (70, 60%) at family level, respectively. Sensitivity analysis based on the mean decrease accuracy (MDA) revealed that the modified catchment area variable is the most relative important variable among the selected variables. Generally, by using feature selection innovative approach and effective learning algorithms, the spatial distribution of soil maps could be made even in low relief lands with acceptable accuracy.