MLP network

The Corona virus epidemic has had numerous detrimental repercussions on the tourism industry. Domestic production and trade have been impeded by the global epidemic and the mitigation efforts implemented when the disease broke out, according to the World Health Organization in particular. The current study looks at how neural networks and soft computing techniques can help provide accurate and effective predictions for the tourism industry, preventing future losses. The study will compare a number of strategies and procedures in order to determine the most successful technology in the tourism industry.

Data This paper presents 3 NN case studies. In the first, fracture toughness of wood was predicted using an expanded MLP network from experimentally measured crack angle, stifhess, density and moisture content. The data is characterized by noise but the model produced physically meaningful nonlinear trends with an R2 value of 0.67. In the second study, hydraulic conductivity (K dday) was estimated fiom ground water solute concentration data collected for a range of K values. Four separate NN needed to be developed for four sub-ranges of K to reduce error. In order to determine the appropriate range of K for a particular system, concentration data were clustered into 4 groups using SOM. The hybrid-model was applied to an experimental aquifer and only 10% difference was found between experimental and NN estimations of K. In the third study, digtal images of lamb chops were used to collect values for 102 geometric and textural variables for meat grading. Principal Component analysis reduced the variables to twelve. Three-and 4layer MLP networks and Discriminant Function Analysis @FA) were performed on the data and the classification accuracy from 3-layer MLP was 83% and was 12% better than that from DFA.

This study describes the components of available forage in Vahregan watershed, Central Iran and highlight issues relating to forage endowment and environmental dynamics. In this study, proper use factor and palatability models and their components use to develop the model of available forage and measure it. The components climatic, topography, land systems, vegetation, land use and grass and shrub species composition changes were analyzed using GIS. In this study source of information are herders, land and livestock owners, research institution and personal field inspections. Responses to the questionnaires were obtained from local government offices and research on rangelands and analyzed in parallel with the results of detailed interviews with pastoralists. The results of the completed overall model of available forage showed that of the 18346.2 hectares in the range area only 8.36% of the rangeland is in good condition and the rest are in fair (13.8%), poor and very poor (77.82%). About 16812.85 ha (91.64%) of the rangeland area shows a downward trend, only 1533.35 ha (8.36%) upward trend. About half of the rangeland is not favorable for domestic livestock grazing. The range condition situation and its trend in consideration of soil and slope properties in this study indicates that the rangeland in Vahregan is a fragile production system, sensitive to soil erosion and rangeland degradation, so for a long time sustainable exploitation should be goes to minimize land degradation in the future, proper management and sustainable exploitation should be implemented in the future

Data This paper presents 3 NN case studies. In the first, fracture toughness of wood was predicted using an expanded MLP network from experimentally measured crack angle, stifhess, density and moisture content. The data is characterized by noise but the model produced physically meaningful nonlinear trends with an R2 value of 0.67. In the second study, hydraulic conductivity (K dday) was estimated fiom ground water solute concentration data collected for a range of K values. Four separate NN needed to be developed for four sub-ranges of K to reduce error. In order to determine the appropriate range of K for a particular system, concentration data were clustered into 4 groups using SOM. The hybrid-model was applied to an experimental aquifer and only 10% difference was found between experimental and NN estimations of K. In the third study, digtal images of lamb chops were used to collect values for 102 geometric and textural variables for meat grading. Principal Component analysis reduced the variables to twelve. Three-and 4layer MLP networks and Discriminant Function Analysis @FA) were performed on the data and the classification accuracy from 3-layer MLP was 83% and was 12% better than that from DFA.

This paper proposes the deterministic generation of auxiliary variables, which outline the seasonal, cyclic and trend components of the time series associated with tourism demand for the machine learning models. To test the contribution of the deterministically generated auxiliary variables, we have employed multilayer perceptron (MLP) regression, and support vector regression (SVR) models, which are the well-known state-of- art machine learning models. These models are used to make multivariate tourism forecasting for Turkey respected to two data sets: raw data set and data set with deterministically generated auxiliary variables. The forecasting performances are compared regards to these two data sets. In terms of relative absolute error (RAE) and root relative squared error (RRSE) measurements, the proposed machine learning models have achieved significantly better forecasting accuracy when the auxiliary variables have been employed.

The face of the earth is always changing due to human activities and natural phenomena. Therefore, to optimize the management of the natural areas, knowledge of the trend, extent and estimation of land cover / use changes is considered necessary. Reviewing these changes through satellite images and evaluating their potential through modeling can help environmental planners and natural resource managers to make more informed decisions. In the present study, quantitative detection and evaluation of changes in vegetation were performed in the areas with combat desertification projects, Shahdad, Bam and Garmsar in Iran, during a 30-year period within 1987, 2002 and 2017. The Normalized Difference Vegetation Index (NDVI) and land use maps were produced using the Enhanced Thematic Mapper Plus (ETM+), Thematic Mapper (TM) and Operational Land Imager (OLI) satellite images in the three corresponding periods for the vegetation/non-vegetation, and agricultural lands. The Kappa coefficient of 0.83 to 0.86, 0.91 to 0.92, and 0.94 to 0.95 was calculated for 1987, 2002, and 2017 respectively, and the total accuracy was between 88 and 97. After providing the land use maps in different years, the monitoring of land use changes was investigated using the Change Detection method. According to the trend of changes during the periods, the results exhibited that the vegetated lands in these three areas had an increasing trend in average 31.33%, and the non-vegetated lands were turned to vegetated lands over time. In other words, they have declined by an average of 35%. Moreover, an increasing trend was found for the agricultural lands during the periods in average 4%. Eventually, the cost-effectiveness of projects implemented in the studied areas was calculated.

Machine learning methods originated from artificial intelligence and today are applied in several fields concerning environmental sciences. Thanks to their powerful nonlinear modelling capability, machine learning methods today are utilized in satellite data processing, general circulation models(GCM), weather and climate prediction, air quality forecasting, analysis and modelling of environmental data, oceanographic and hydrological forecasting, ecological modelling, and monitoring of snow, ice and forests. Currently, the popularity of neural networks is growing; their areas of application are constantly expanding. In these conditions, the task of choosing a convenient tool for utilizing in environmental science with neural networks becomes urgent. There are many tools for working with neural networks, but each of them has its own drawbacks. So most of the existing tools require users to have programming knowledge; there are no tools to help quickly select the optimal network structure for the problem being solved. The purpose of the research is to simplify the process of choosing the optimal structure of an artificial neural network by developing an application with a graphical user interface with a visual representation of the stages of creating and learning neural networks in environmental sciences. The object of research is artificial feed-forward neural networks. Research work on the study, comparison and analysis of existing tools for the creation, learning and use of artificial neural networks has been carried out. Based on the research results, an application with a graphical interface aimed at solving the assigned tasks has been developed. An application developed to achieve this goal works correctly, without failures, and allows creating and learning feed-forward neural networks without programming knowledge.

The face of the earth is always changing due to human activities and natural phenomena. Therefore, to optimize the management of the natural areas, knowledge of the trend, extent and estimation of land cover / use changes is considered necessary. Reviewing these changes through satellite images and evaluating their potential through modeling can help environmental planners and natural resource managers to make more informed decisions. In the present study, quantitative detection and evaluation of changes in vegetation were performed in the areas with combat desertification projects, Shahdad, Bam and Garmsar in Iran, during a 30-year period within 1987, 2002 and 2017. The Normalized Difference Vegetation Index (NDVI) and land use maps were produced using the Enhanced Thematic Mapper Plus (ETM+), Thematic Mapper (TM) and Operational Land Imager (OLI) satellite images in the three corresponding periods for the vegetation/non-vegetation, and agricultural lands. The Kappa coefficient of 0.83 to 0.86, 0.91 to 0.92, and 0.94 to 0.95 was calculated for 1987, 2002, and 2017 respectively, and the total accuracy was between 88 and 97. After providing the land use maps in different years, the monitoring of land use changes was investigated using the Change Detection method. According to the trend of changes during the periods, the results exhibited that the vegetated lands in these three areas had an increasing trend in average 31.33%, and the non-vegetated lands were turned to vegetated lands over time. In other words, they have declined by an average of 35%. Moreover, an increasing trend was found for the agricultural lands during the periods in average 4%. Eventually, the cost-effectiveness of projects implemented in the studied areas was calculated.

Rangelands production measurement is time-consuming and expensive. Therefore, models are often employed to simulate rangelands conditions as a supplement. Artificial neural network (ANN) is widely used for modeling in environmental studies, yet it cannot preset its results in the form of a map or geo-referenced data. We used ANN to estimate the spatial distribution of rangelands production, then a geographic information system (GIS) was applied as a pre-processing and post-processing framework in rangelands production modeling. The ANN was trained (Rsqr = 0.95, MSE = 0.02) and tested using data from the Baladeh rangelands located in the northern part of Iran. Rangelands production was simulated using a multi-layer perceptron (MLP) network. We estimated rangelands production (using many plots and field studies) as the network output, along with the influencing factors in the production (vegetation, climatic, topographic, edaphic and human factors) as the inputs. After modeling and model optimizing in ANN, the model test was performed (Rsqr=0.8, MSE=0.3). Furthermore, the studied area was divided with the pixels 100×100 m (raster format) in the GIS medium. Then, the digital layers of the network inputs were combined and a raster layer was prepared including the network inputs values and geographic coordinate. The values of pixels (network inputs) were imported in ANN (NeuroSolutions software). Rangelands production was simulated using the validated optimum network in the sites without production measurements. In the next step, the results of ANN simulation were imported in the GIS medium, then rangelands production map was prepared based on the estimated results of ANN. The results indicated that integrating ANN and GIS exhibits high accuracy and performance in rangelands production estimation. Hence, the prepared rangelands production map can be used for planning and managing the rangelands.

This study presents an extension of the Gaussian process regression model for multiple-input multiple-output forecasting. This approach allows modelling the cross-dependencies between a given set of input variables and generating a vectorial prediction. Making use of the existing correlations in international tourism demand to all seventeen regions of Spain, the performance of the proposed model is assessed in a multiple-step-ahead forecasting comparison. The results of the experiment in a multivariate setting show that the Gaussian process regression model significantly improves the forecasting accuracy of a multi-layer perceptron neural network used as a benchmark. The results reveal that incorporating the connections between different markets in the modelling process may prove very useful to refine predictions at a regional level.