scholarly journals The use of machine learning algorithms to design a generalized simplified denitrification model

2010 ◽  
Vol 7 (2) ◽  
pp. 2313-2360 ◽  
Author(s):  
F. Oehler ◽  
J. C. Rutherford ◽  
G. Coco
Keyword(s):  
Machine Learning ◽  
Water Content ◽  
Soil Water ◽  
Soil Water Content ◽  
Nitrate Concentration ◽  
Learning Algorithms ◽  
Machine Learning Algorithms ◽  
Boosted Regression Trees ◽  
Nitrogen Emissions ◽  
Denitrification Potential

Abstract. We designed generalized simplified models using machine learning algorithms (ML) to assess denitrification at the catchment scale. In particular, we designed an artificial neural network (ANN) to simulate total nitrogen emissions from the denitrification process. Boosted regression trees (BRT, another ML) was also used to analyse the relationships and the relative influences of different input variables towards total denitrification. To calibrate the ANN and BRT models, we used a large database obtained by collating datasets from the literature. We developed a simple methodology to give confidence intervals for the calibration and validation process. Both ML algorithms clearly outperformed a commonly used simplified model of nitrogen emissions, NEMIS. NEMIS is based on denitrification potential, temperature, soil water content and nitrate concentration. The ML models used soil organic matter % in place of a denitrification potential and pH as a fifth input variable. The BRT analysis reaffirms the importance of temperature, soil water content and nitrate concentration. Generality of the ANN model may also be improved if pH is used to differentiate between soil types. Further improvements in model performance can be achieved by lessening dataset effects.

scholarly journals The use of machine learning algorithms to design a generalized simplified denitrification model

2010 ◽  
Vol 7 (10) ◽  
pp. 3311-3332 ◽  
Author(s):  
F. Oehler ◽  
J. C. Rutherford ◽  
G. Coco
Keyword(s):  
Machine Learning ◽  
Water Content ◽  
Soil Water ◽  
Soil Water Content ◽  
Nitrate Concentration ◽  
Machine Learning Algorithms ◽  
Boosted Regression Trees ◽  
Simplified Model ◽  
Nitrogen Emissions ◽  
Denitrification Potential

Abstract. We propose to use machine learning (ML) algorithms to design a simplified denitrification model. Boosted regression trees (BRT) and artificial neural networks (ANN) were used to analyse the relationships and the relative influences of different input variables towards total denitrification, and an ANN was designed as a simplified model to simulate total nitrogen emissions from the denitrification process. To calibrate the BRT and ANN models and test this method, we used a database obtained collating datasets from the literature. We used bootstrapping to compute confidence intervals for the calibration and validation process. Both ML algorithms clearly outperformed a commonly used simplified model of nitrogen emissions, NEMIS, which is based on denitrification potential, temperature, soil water content and nitrate concentration. The ML models used soil organic matter % in place of a denitrification potential and pH as a fifth input variable. The BRT analysis reaffirms the importance of temperature, soil water content and nitrate concentration. Generalization, although limited to the data space of the database used to build the ML models, could be improved if pH is used to differentiate between soil types. Further improvements in model performance and generalization could be achieved by adding more data.

scholarly journals Skin Characterizations by Using Contact Capacitive Imaging and High-Resolution Ultrasound Imaging with Machine Learning Algorithms

2021 ◽  
Vol 11 (18) ◽  
pp. 8714
Author(s):  
Elena Chirikhina ◽  
Andrey Chirikhin ◽  
Sabina Dewsbury-Ennis ◽  
Francesco Bianconi ◽  
Perry Xiao
Keyword(s):  
Machine Learning ◽  
High Resolution ◽  
Stratum Corneum ◽  
Image Classification ◽  
Water Content ◽  
Ultrasound Imaging ◽  
Skin Diseases ◽  
Learning Algorithms ◽  
Machine Learning Algorithms ◽  
Body Parts

We present our latest research on skin characterizations by using Contact Capacitive Imaging and High-Resolution Ultrasound Imaging with Machine Learning algorithms. Contact Capacitive Imaging is a novel imaging technology based on the dielectric constant measurement principle, with which we have studied the skin water content of different skin sites and performed image classification by using pre-trained Deep Learning Neural Networks through Transfer Learning. The results show lips and nose have the lowest water content, whilst cheek, eye corner and under-eye have the highest water content. The classification yields up to 83.8% accuracy. High-Resolution Ultrasound Imaging is a state-of-the-art ultrasound technology, and can produce high-resolution images of the skin and superficial soft tissue to a vertical resolution of about 40 microns, with which we have studied the thickness of different skin layers, such as stratum corneum, epidermis and dermis, around different locations on the face and around different body parts. The results show the chin has the highest stratum corneum thickness, and the arm has the lowest stratum corneum thickness. We have also developed two feature-based image classification methods which yield promising results. The outcomes of this study could provide valuable guidelines for cosmetic/medical research, and methods developed in this study can also be extended for studying damaged skin or skin diseases. The combination of Contact Capacitive Imaging and High-Resolution Ultrasound Imaging could be a powerful tool for skin studies.

scholarly journals Surface Motion Prediction and Mapping for Road Infrastructures Management by PS-InSAR Measurements and Machine Learning Algorithms

2020 ◽  
Vol 12 (23) ◽  
pp. 3976
Author(s):  
Nicholas Fiorentini ◽  
Mehdi Maboudi ◽  
Pietro Leandri ◽  
Massimo Losa ◽  
Markus Gerke
Keyword(s):  
Machine Learning ◽  
Learning Algorithms ◽  
Central Italy ◽  
Machine Learning Algorithms ◽  
Boosted Regression Trees ◽  
Bayesian Optimization ◽  
Support Vector ◽  
Surface Motion ◽  
Persistent Scatterers ◽  
Taylor Diagram

This paper introduces a methodology for predicting and mapping surface motion beneath road pavement structures caused by environmental factors. Persistent Scatterer Interferometric Synthetic Aperture Radar (PS-InSAR) measurements, geospatial analyses, and Machine Learning Algorithms (MLAs) are employed for achieving the purpose. Two single learners, i.e., Regression Tree (RT) and Support Vector Machine (SVM), and two ensemble learners, i.e., Boosted Regression Trees (BRT) and Random Forest (RF) are utilized for estimating the surface motion ratio in terms of mm/year over the Province of Pistoia (Tuscany Region, central Italy, 964 km2), in which strong subsidence phenomena have occurred. The interferometric process of 210 Sentinel-1 images from 2014 to 2019 allows exploiting the average displacements of 52,257 Persistent Scatterers as output targets to predict. A set of 29 environmental-related factors are preprocessed by SAGA-GIS, version 2.3.2, and ESRI ArcGIS, version 10.5, and employed as input features. Once the dataset has been prepared, three wrapper feature selection approaches (backward, forward, and bi-directional) are used for recognizing the set of most relevant features to be used in the modeling. A random splitting of the dataset in 70% and 30% is implemented to identify the training and test set. Through a Bayesian Optimization Algorithm (BOA) and a 10-Fold Cross-Validation (CV), the algorithms are trained and validated. Therefore, the Predictive Performance of MLAs is evaluated and compared by plotting the Taylor Diagram. Outcomes show that SVM and BRT are the most suitable algorithms; in the test phase, BRT has the highest Correlation Coefficient (0.96) and the lowest Root Mean Square Error (0.44 mm/year), while the SVM has the lowest difference between the standard deviation of its predictions (2.05 mm/year) and that of the reference samples (2.09 mm/year). Finally, algorithms are used for mapping surface motion over the study area. We propose three case studies on critical stretches of two-lane rural roads for evaluating the reliability of the procedure. Road authorities could consider the proposed methodology for their monitoring, management, and planning activities.

scholarly journals Machine Learning Approaches to Develop Pedotransfer Functions for Tropical Sri Lankan Soils

Water ◽  
2019 ◽  
Vol 11 (9) ◽  
pp. 1940 ◽  
Author(s):  
M.H.J.P. Gunarathna ◽  
Kazuhito Sakai ◽  
Tamotsu Nakandakari ◽  
Kazuro Momii ◽  
M.K.N. Kumari
Keyword(s):  
Machine Learning ◽  
Water Content ◽  
Learning Algorithms ◽  
Field Capacity ◽  
Machine Learning Algorithms ◽  
Pedotransfer Functions ◽  
Sri Lankan ◽  
Learning Approaches ◽  
Tropical Regions ◽  
Wilting Point

Poor data availability on soil hydraulic properties in tropical regions hampers many studies, including crop and environmental modeling. The high cost and effort of measurement and the increasing demand for such data have driven researchers to search for alternative approaches. Pedotransfer functions (PTFs) are predictive functions used to estimate soil properties by easily measurable soil parameters. PTFs are popular in temperate regions, but few attempts have been made to develop PTFs in tropical regions. Regression approaches are widely used to develop PTFs worldwide, and recently a few attempts were made using machine learning methods. PTFs for tropical Sri Lankan soils have already been developed using classical multiple linear regression approaches. However, no attempts were made to use machine learning approaches. This study aimed to determine the applicability of machine learning algorithms in developing PTFs for tropical Sri Lankan soils. We tested three machine learning algorithms (artificial neural networks (ANN), k-nearest neighbor (KNN), and random forest (RF)) with three different input combination (sand, silt, and clay (SSC) percentages; SSC and bulk density (BD); SSC, BD, and organic carbon (OC)) to estimate volumetric water content (VWC) at −10 kPa, −33 kPa (representing field capacity (FC); however, most studies in Sri Lanka use −33 kPa as the FC) and −1500 kPa (representing the permanent wilting point (PWP)) of Sri Lankan soils. This analysis used the open-source data mining software in the Waikato Environment for Knowledge Analysis. Using a wrapper approach and best-first search method, we selected the most appropriate inputs to develop PTFs using different machine learning algorithms and input levels. We developed PTFs to estimate FC and PWP and compared them with the previously reported PTFs for tropical Sri Lankan soils. We found that RF was the best algorithm to develop PTFs for tropical Sri Lankan soils. We tried to further the development of PTFs by adding volumetric water content at −10 kPa as an input variable because it is quite an easily measurable parameter compared to the other targeted VWCs. With the addition of VWC at −10 kPa, all machine learning algorithms boosted the performance. However, RF was the best. We studied the functionality of finetuned PTFs and found that they can estimate the available water content of Sri Lankan soils as well as measurements-based calculations. We identified RF as a robust alternative to linear regression methods in developing PTFs to estimate field capacity and the permanent wilting point of tropical Sri Lankan soils. With those findings, we recommended that PTFs be developed using the RF algorithm in the related software to make up for the data gaps present in tropical regions.

A review of pressure strain correlation modeling for Reynolds stress models

2019 ◽  
Vol 234 (8) ◽  
pp. 1528-1544 ◽  
Author(s):  
JP Panda
Keyword(s):  
Machine Learning ◽  
Turbulent Flows ◽  
Reynolds Stress ◽  
Learning Algorithms ◽  
Machine Learning Algorithms ◽  
Boosted Regression Trees ◽  
Modeling Framework ◽  
Stress Modeling ◽  
Flow Evolution ◽  
Stress Models

Most investigations of turbulent flows in academic studies and industrial applications use turbulence models. Out of the different turbulence modeling approaches Reynolds stress models have the highest potential to replicate complex turbulent flow phenomena at a reasonable computational expense. The Reynolds stress modeling framework is constituted by individual closures that approximate the effects of separate turbulence processes like dissipation, turbulent transport, pressure strain correlation, etc. Owing to its complexity and importance in flow evolution the modeling of the pressure strain correlation mechanism is considered the crucial challenge for the Reynolds stress modeling framework. In the present work, the modeling of the pressure strain correlation for homogeneous turbulent flows is reviewed. The importance of the pressure strain correlation and its effects on flow evolution via energy transfer are established. The fundamental challenges in pressure stain correlation modeling are analyzed and discussed. Starting from the governing equations we outline the theory behind models for both the slow and rapid pressure strain correlation. Established models for both these are introduced and their successes and shortcomings are illustrated using theoretical analysis, computational fluid dynamics simulations, and comparisons against experimental and numerical studies. Recent advances and developments in this field are presented and discussed. The application of machine learning algorithms such as Deep Neural Networks, Random Forests, and Gradient Boosted Regression Trees is summarized and examined. We report fundamental problems in the application of machine learning algorithms for pressure strain correlation modeling. Finally, challenges and hurdles for pressure strain correlation modeling are outlined and explained in detail to guide future investigations.

Sign in / Sign up

Export Citation Format

Share Document