关键词： Food authentication   Organic apple   Diffraction grating   Machine learning   Chemometrics  

摘要： Optical measuring technologies coupled with machine learning algorithms can be used to build a home-made sensor system. We built such a sensor system using a smartphone and a diffraction grating sheet. Diffraction images were captured under white light illumination and converted into a data matrix for data analysis. In this paper we present a systematic evaluation of this sensor system on the task of differentiating organic apples from conventional ones. We used the sensor system to measure 150 organic and conventional apples as rainbow images. We processed the rainbow images using computer vision techniques, built machine learning and chemometrics models, and used the resultant models to classify testing samples. Moreover, a comparative study was conducted where the same set of apples were scanned by a commercial spectrometer resulting in spectral data of the apple samples and classification was undertaken using partial least squares discriminant analysis (PLS-DA). Experimental results show that state of the art machine learning algorithms such as support vector machine (SVM) and locally weighted partial least squares classifier (LW-PLSC) are effective in handling low-quality image data with classification accuracies of 93 - 100%. These results suggest that the sensor system is convenient and low-cost, and provides a fast, effective, non-destructive and viable solution for in-line food authentication.

关键词： Convolutional neural networks   Feature extraction   Machine learning   Prediction   Classification   Aesthetics assessment   Hybrid model   Transfer learning  

摘要： The automatic assessment of the aesthetic value of an image is a task with many applications but really complex and challenging, due to the subjective component of the aesthetics for humans. The computational systems that carry out this task are usually composed of a set of ad hoc metrics proposed by the researchers and a machine learning system. We propose a new approach that fully automates the metrics creation process, its filtering and adjustment without human subjectivity. Thus, it does not depend on the authors' human aesthetic intuitions. Our proposal is therefore based on the integration of two machine learning algorithms: CNN, which works as a feature extractor, and Correlation by Genetic Search (CGS)-a novel regression method, working as a supervised learning method. CGS is based on the creation of an adjusted linear regression model using Pearson's correlation as a measure of performance in an evolutionary process. Experiments were conducted on a very well-known aesthetics database called "***" with more than a million images from over 400,000 users. The comparison of results with other approaches using the same dataset demonstrates that the fusion of CNN transfer learning features with this specific machine learning method has achieved robust and significantly better results than other state-of-the-art methods and hybrid approaches in terms of AUROC (0.93), accuracy (0.93) and Pearson's correlation value (0.94).

关键词： crop residue cover   crop residue indices   empirical regression   machine learning regression   Sentinel-2 MSI   textural feature   unmanned aerial vehicle  

摘要： Quantifying crop residue cover (CRC) on field surfaces is important for monitoring the tillage intensity and promoting sustainable management. Remote-sensing-based techniques have proven practical for determining CRC, however, the methods used are primarily limited to empirical regression based on crop residue indices (CRIs). This study provides a systematic evaluation of empirical regressions and machine learning (ML) algorithms based on their ability to estimate CRC using Sentinel-2 Multispectral Instrument (MSI) data. Unmanned aerial vehicle orthomosaics were used to extracted ground CRC for training Sentinel-2 data-based CRC models. For empirical regression, nine MSI bands, 10 published CRIs, three proposed CRIs, and four mean textural features were evaluated using univariate linear regression. The best performance was obtained by a three-band index calculated using (B2 - B4)/(B2 - B12), with an R-cv(2) of 0.63 and RMSEcv of 6.509%, using a 10-fold cross-validation. The methodologies of partial least squares regression (PLSR), artificial neural network (ANN), Gaussian process regression (GPR), support vector regression (SVR), and random forest (RF) were compared with four groups of predictors, including nine MSI bands, 13 CRIs, a combination of MSI bands and mean textural features, and a combination of CRIs and textural features. In general, ML approaches achieved high accuracy. A PLSR model with 13 CRIs and textural features resulted in an accuracy of R-cv(2) = 0.66 and RMSEcv = 6.427%. An RF model with predictors of MSI bands and textural features estimated CRC with an R-cv(2) = 0.61 and RMSEcv = 6.415%. The estimation was improved by an SVR model with the same input predictors (R-cv(2) = 0.67, RMSEcv = 6.343%), followed by a GPR model based on CRIs and textural features. The performance of GPR models was further improved by optimal input variables. A GPR model with six input variables, three MSI bands and three textural features, performed the best, with R-cv

关键词： Ultrasound   Nondestructive evaluation   Self-organizing map   Bag of features   Radial basis function  

摘要： In this study, we propose a new method for the nondestructive measurement of the fat content and texture of fish meat using machine learning and the bag of features approach. We employed two machine learning methods, that is, a self-organizing map (SOM) and radial basis function (RBF) network. The SOM was applied to symbolize the pattern of the frequency spectrum extracted from ultrasound signals and to generate key features for the bag of features technique. The RBF network was applied to estimate the fat content and texture of fish meat from the bag of features histogram. We verified the accuracy of the fat content and texture estimations given by the proposed method through a series of experiments. The results showed that the fat content and texture of fish meat was estimated more accurately using the proposed method than by the conventional approach.

关键词： Algorithm selection   Digital soil mapping   Machine learning   Model evaluation  

摘要： Digital soil mapping (DSM) increasingly makes use of machine learning algorithms to identify relationships between soil properties and multiple covariates that can be detected across landscapes. Selecting the appropriate algorithm for model building is critical for optimizing results in the context of the available data. Over the past decade, many studies have tested different machine learning (ML) approaches on a variety of soil data sets. Here, we review the application of some of the most popular ML algorithms for digital soil mapping. Specifically, we compare the strengths and weaknesses of multiple linear regression (MLR), k-nearest neighbors (KNN), support vector regression (SVR), Cubist, random forest (RF), and artificial neural networks (ANN) for DSM. These algorithms were compared on the basis of five factors: (1) quantity of hyperparameters, (2) sample size, (3) covariate selection, (4) learning time, and (5) interpretability of the resulting model. If training time is a limitation, then algorithms that have fewer model parameters and hyperparameters should be considered, e.g., MLR, KNN, SVR, and Cubist. If the data set is large (thousands of samples) and computation time is not an issue, ANN would likely produce the best results. If the data set is small (<100), then Cubist, KNN, RF, and SVR are likely to perform better than ANN and MLR. The uncertainty in predictions produced by Cubist, KNN, RF, and SVR may not decrease with large datasets. When interpretability of the resulting model is important to the user, Cubist, MLR, and RF are more appropriate algorithms as they do not function as "black boxes." There is no one correct approach to produce models for predicting the spatial distribution of soil properties. Nonetheless, some algorithms are more appropriate than others considering the nature of the data and purpose of mapping activity. (C) 2019 Elsevier Inc. All rights reserved.

关键词： Ohmic heating   Minas frescal cheese   Sensory drivers of liking   Consumer acceptance   Machine learning methods  

摘要： The consumer acceptance (n = 100) and the sensory drivers of liking of Minas frescal cheese manufactured with milk subjected to ohmic heating (0, 4, 8, and 12 V/cm(-1), CONY, OH4, OH8, and OH12, 72-75 degrees C/15 s) were investigated. Machine learning techniques (random forest, gradient boosted trees, and extreme learning machine;RF, GBT, and ELM) were used to determine the sensory drivers of liking. No significant differences were observed among the cheeses for most of the sensory attributes, for all treatments, suggesting that ohmic heating may be an adequate technology for Minas Frescal cheese processing with the advantage of improving its overall liking. Machine learning methods presented a good agreement with the experimental data, allowing the identification of the attribute's juiciness, white color, homogenous mass, Minas Frescal cheese flavor as the sensory drivers of liking, while the attribute bitter taste was identified as a driver of disliking. These results should be taken into consideration when adopting emerging technologies, such as ohmic heating for the manufacture of Minas frescal cheese.

关键词： Classification   fractal dimension   Hurst exponent   attributes   training   testing   metrics   algorithms   anomalies   attacks  

摘要： In this paper, it is proposed to improve the efficiency of binary classification of network traffic anomalous behavior by introducing an additional informative feature -fractal dimension. The overall effectiveness of the proposed method is estimated by evaluating the quality of binary classification using the algorithmsDecision Tree Classifier,Random ForestandAda Booston the example of using the NSL-KDD database. It is shown that adding the fractal dimension in the binary classification of attacks, gives improvement of theprecisionmetric in average by 6%, and for AUC-ROC about 10% for all considered classification algorithms. Furthermore, introduction of fractal dimension as an additional feature has allowed to significantly reduce the time of training and testing of binary classification. So, for theRandom Forestalgorithm, the decrease in processing time was more than 3 times, and for theDecision Tree Classifiermore than 2 times.

关键词： pumping tests   machine learning   time-series analysis   cluster analysis   random forest method  

摘要： Pumping tests are very important means for investigating aquifer properties;however, interpreting the data using common analytical solutions become invalid in complex aquifer systems. The paper aims to explore the potential of machine learning methods in retrieving the pumping tests information in a field site in the Democratic Republic of Congo. A newly planned mining site with a pumping test of three pumping wells and 28 observation wells over one month was chosen to analyze the significance of machine learning methods in the pumping test analysis. Widely used machine learning methods, including correlation, cluster, time-series analysis, artificial neural network (ANN), support vector machine (SVR), random forest (RF) method, and linear regression, are all used in this study. Correlation and cluster analyses among wells provide visual pictures of possible hydraulic connections. The pathway with the best permeability ranges from the depth of 250 m to 350 m. Time-series analysis perfectly captured changes of drawdowns within the three pumping wells. The RF method is found to have the higher accuracy and the lower sensitivity to model parameters than ANN and SVR methods. The coupling of the linear regressive model and analytical solutions is applied to estimate hydraulic conductivities. The results found that ML methods can significantly and effectively improve our understanding of pumping tests by revealing inherent information hidden in those tests.

关键词： Toxicology   Structure-activity relationships   machine learning   mitochondrial toxicity   structural alerts  

摘要： Over the last few years more and more organ and idiosyncratic toxicities were linked to mitochondrial toxicity. Despite well-established assays, such as the seahorse and Glucose/Galactose assay, an in silico approach to mitochondrial toxicity is still feasible, particularly when it comes to the assessment of large compound libraries. Therefore, in silico approaches could be very beneficial to indicate hazards early in the drug development pipeline. By combining multiple endpoints, we derived the largest so far published dataset on mitochondrial toxicity. A thorough data analysis shows that molecules causing mitochondrial toxicity can be distinguished by physicochemical properties. Finally, the combination of machine learning and structural alerts highlights the suitability for in silico risk assessment of mitochondrial toxicity.

关键词： Soil moisture   Machine learning   Data assimilation   Physically-based model   Artificial neural network  

摘要： Soil moisture plays a critical role as an essential component of the global water resources by regulating mass and energy exchange between land surface and atmosphere. Quantification of these exchange processes requires accurate characterization and simulation of soil water movement. Physically-based models (PBMs) and machine learning methods (MLMs) can both be used in soil moisture simulation. However, their performances in soil water simulation have only been compared in a limited number of cases. Moreover, almost all of them are conducted in field studies each with fixed soil, initial condition, and boundary condition. Here, we developed three artificial neural network (ANN) frameworks, and made clearer and more systematic comparisons between them and a PBM-Ross numerical model solving Richards equation and parameter estimation using a data assimilation approach (iterative ensemble smoother, Ross-IES) in synthetic and real-world conditions. Compared with the ANNs, Ross-IES is more significantly affected by physical model uncertainties such as soil heterogeneity, initial and boundary conditions, while both methods are affected by observation noise. For Ross-IES, the errors from boundary conditions and hydraulic parameter conceptualization are found to be more prominent than that of observation noise and therefore are suggested to be identified first. Meanwhile, the ANNs have difficulty in simulating the peaks and troughs of the soil water time series as well as in situations where the soil moisture is constantly saturated. ANNs yield a superior simulation when the nonlinear relationship between the response variables and driving data is weak, while the performance of Ross-IES is governed by the prior soil hydraulic information. In addition, Ross-IES approach requires much higher computational cost than the ANNs. ANN-MS performs best among the three ANN-based machine learning models and demonstrates great data mining ability and robustness against overfitting.