Machine Learning Toolbox Research Articles

Prediction and Detection of Ventricular Fibrillation Using Complex Features and AI-Based Classification

We analyzed the possibility of detecting and predicting ventricular fibrillation (VF), a medical emergency that may put people’s lives at risk, as the medical prognosis depends on the time in which medical personnel intervene. Therefore, besides immediate detection of VF, the possibility of predicting VF 40 or even 50 min in advance was analyzed. For testing the proposed algorithm, we used ECG signals taken from the MIT-BIH database, namely, Malignant Ventricular Ectopy Database, Sudden Cardiac Death Holter Database and Normal Sinus Rhythm Database. The presented method is based on features extracted from the ECG signals in the time domain, frequency domain and complexity measures. For VF detection, the possibility of identifying the VF episode in the first 3 s after its occurrence was tested. For this, the first 3 s immediately after the appearance of VF were cut out and the features were computed on these sections. For VF prediction, 3 min of the ECG signal clipped 40 or 50 min before VF onset was used. Then, on these pieces of ECG signal, the specific features were calculated for 1 s segments. For the normal signal situation, 3 min was randomly selected from the database with normal ECGs. For the classification or detection stage, both an MLP-type neural network and the classifiers from the Machine Learning toolbox of the MATLAB® environment were used. The results obtained for both detection and classification are over 94% in both cases. The novelty of our results compared to those previously obtained is the time interval with which the possibility of prediction was analyzed, namely, 50 min in advance of the VF installation date. This means that the patient will be informed that it is possible to suffer a VF and has time to take the necessary measures to overcome a possible medical emergency.

A machine learning toolbox for the analysis of sharp-wave ripples reveals common waveform features across species

The study of sharp-wave ripples has advanced our understanding of memory function, and their alteration in neurological conditions such as epilepsy is considered a biomarker of dysfunction. Sharp-wave ripples exhibit diverse waveforms and properties that cannot be fully characterized by spectral methods alone. Here, we describe a toolbox of machine-learning models for automatic detection and analysis of these events. The machine-learning architectures, which resulted from a crowdsourced hackathon, are able to capture a wealth of ripple features recorded in the dorsal hippocampus of mice across awake and sleep conditions. When applied to data from the macaque hippocampus, these models are able to generalize detection and reveal shared properties across species. We hereby provide a user-friendly open-source toolbox for model use and extension, which can help to accelerate and standardize analysis of sharp-wave ripples, lowering the threshold for its adoption in biomedical applications.

ТЕОРЕТИЧЕСКИЕ ПРЕДПОСЫЛКИ ОПРЕДЕЛЕНИЯ ФОРМЫ И РАЗМЕРА ОЧАГА РАСПРОСТРАНЕНИЯ ОЗОНА В ЗЕРНОВОМ ВОРОХЕ

It is known that at the present stage of development of the agro-industrial complex, the ozonation process is increasingly being used. This is an environmentally friendly method of processing agricultural raw materials in order to disinfect, deodorize, disinsect, increase the growth qualities of seeds, etc. It is worth noting that, despite such an active use of ozone, the question remains not fully understood how the gas spreads in the grain during its processing. The article presents the results of research to determine the shape and size of the source of the ozone-air mixture divergence in the grain heap. It has been experimentally established that the ozone in the grain diverges in a shape close to an ellipsoid. The size of the lesion depends on the saturation time and, as a result, the received dose of ozone treatment. As the duration of the experiment increases, the radius of the ellipsoid increases more than its height. Based on the results of experimental studies in the Matlab program, a model of the behavior of ozone in grain is constructed depending on its processing modes. Using regressions based on Gauss processes, the shape and size of the distribution center of the ozone-air mixture in the grain were predicted with a change in the ozonation time. The selection of model parameters is based on Bayesian optimization in the Regression Learner application from the Statistics and Machine Learning Toolbox of the Matlab application software package. The electronic program allows you to optimize hyperparameters and obtain confidence intervals for predicted values. The research results make it possible to predict the behavior of ozone in grain, which is necessary to determine the number of gas supply nozzles inside a granary of any volume.

Nonlinear Model Predictive Control-based Collision Avoidance for Mobile Robot

This work proposes an efficient and safe single-layer Nonlinear Model Predictive Control (NMPC) system based on LiDAR to solve the problem of autonomous navigation in cluttered environments with previously unidentified static and dynamic obstacles of any shape. Initially, LiDAR sensor data is collected. Then, the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm, is used to cluster the (Lidar) points that belong to each obstacle together. Moreover, a Minimum Euclidean Distance (MED) between the robot and each obstacle with the aid of a safety margin is utilized to implement safety-critical obstacle avoidance rather than existing methods in the literature that depend on enclosing the obstacles with a circle or minimum bounding ellipse. After that, to impose avoidance constraints with feasibility guarantees and without compromising stability, an NMPC for set-point stabilization is taken into consideration with a design strategy based on terminal inequality and equality constraints. Consequently, numerous obstacles can be avoided at the same time efficiently and rapidly through unstructured environments with narrow corridors. Finally, a case study with an omnidirectional wheeled mobile robot (OWMR) is presented to assess the proposed NMPC formulation for set-point stabilization. Furthermore, the efficacy of the proposed system is tested by experiments in simulated scenarios using a robot simulator named CoppeliaSim in combination with MATLAB which utilizes the CasADi Toolbox, and Statistics and Machine Learning Toolbox. Two simulation scenarios are considered to show the performance of the proposed framework. The first scenario considers only static obstacles while the second scenario is more challenging and contains static and dynamic obstacles. In both scenarios, the OWMR successfully reached the target pose (1.5m, 1.5m, 0°) with a small deviation. Four performance indices are utilized to evaluate the set-point stabilization performance of the proposed control framework including the steady-state error in the posture vector which is less than 0.02 meters for position and 0.012 for orientation, and the integral of norm squared actual control inputs which is 19.96 and 21.74 for the first and second scenarios respectively. The proposed control framework shows a positive performance in a narrow-cluttered environment with unknown obstacles.

Data-Driven Tunnel Oxide Passivated Contact Solar Cell Performance Analysis Using Machine Learning.

Tunnel oxide passivated contacts (TOPCon) have gained interest as a way to increase the energy conversion efficiency of silicon solar cells, and the International Technology Roadmap of Photovoltaics forecasts TOPCon to become an important technology despite a few remaining challenges. To review the recent development of TOPCon cells, this work has compiled a dataset of all device data found in current literature, which sums up to 405 devices from 131 papers. This may seem like a surprisingly small number of cells given the recent interest in the TOPCon architecture, but it illustrates a problem of data dissemination in the field. Notwithstanding the limited number of cells, there is a great diversity in cell manufacturing procedures, and this work observes a gradual increase in performance indicating that the field has not yet converged on a set of best practices. By analyzing the data using statistical methods and machine learning (ML) algorithms, this work is able to reinforces some commonly held hypotheses related to the performance differences between different device architectures. This work also identifies a few more unintuitive feature combinations that would be of interest for further experimentally studies. This work also aims to inspire improvements in data management and dissemination within the TOPCon community.

A novel machine-learning based prediction model for ship manoeuvring emissions by using bridge simulator

Emissions from ships in ports have primary adverse impacts on human health and the regional environment, as ports are close to residential areas. This study investigates the emissions originated from own ship and tugs assisting the manoeuvres. More than 300,000 row main engine rpm data of the own ship and tugboats used in the berthing manoeuvre to a fictitious port executed by 92 actual maritime pilots in a full mission bridge simulator were exported from the system. Based on the rpm data, emission values were calculated by bottom-up methodologies where resulting emissions are differ up to 1.85 times. A significant correlation between the emission results and pilots’ demographics and experience-based backgrounds are observed. By utilising MATLAB Machine Learning Toolbox, a manoeuvring emission footprint prediction model is developed for the pilots with a consistency of 73%. The proposed model provides a solution that can support strategic planning in determining emission footprints and developing mitigation measures in a cost-effective manner.

Surrogate Modelling of an Industrial Distillation Column Obtained from Statistical Techniques and Machine Learning

Purpose: To study a case of modeling an industrial distillation system, using the Aspen Plus simulator as the mathematical model and subsequently generating surrogate models using Machine Learning techniques in Matlab. Theoretical Framework: Metamodels are reduced models obtained from data generated by rigorous models, replacing them entirely or partially when the computational codes originating from them require excessively large computational effort to be used feasibly. Method/Design/Approach: The simulation of the process was performed in Aspen Plus. Subsequently, the most important variables were selected, and an experimental design was created using Latin Hypercube Sampling in Matlab, generating points to be used in sensitivity analysis in Aspen Plus. The next step was the construction of metamodels using the Statistics and Machine Learning toolbox in Matlab, employing linear regression and Gaussian process regression models. Finally, a statistical analysis was conducted. Results and Conclusion: The distillation system simulation converged, and the obtained metamodels had good regression indicators, especially the Gaussian process regression models, making them the most suitable for representing the rigorous Aspen Plus model. Research Implications: Understanding the multicomponent distillation process integrated with computational tools and data regression models, leading to a reduction in computational effort. Originality/Value: Development of a tool that enables the simulation and evaluation of a distillation process without the need to acquire software with rigorous equation databases.

Employing ANN for daylight and energy prediction of hot climate office buildings: a case study of new Cairo, Egypt

ABSTRACT Indoor lighting conditions are significantly impacted by solar shading. It may be extremely difficult, time-consuming, and expensive to use solar energy efficiently, especially in hot climates. In this paper, the artificial neural network (ANN) is used to forecast and analyze daylight and energy intensity usage for electrical lighting and cooling loads in the early design stages of office buildings. This research will investigate the effect of using different fixed shading systems, including fins, inside louvers, and outside louvers, on an office building in New Cairo, Egypt. There are three steps in the framework of the analysis. Initially, a field experiment is conducted for validation. Then, 3700 simulations on Honeybee for Grasshopper are used to generate a dataset for training, validation, and testing of the ANN model at the third step using the MATLAB® machine learning toolbox. The performance of the ANN model is evaluated using the R-Squared and mean-squared error (MSE) metrics. The resulting ANN models are 446 times quicker than the simulation software and have R-Squared values of 1, 1, 0.98, and 0.983 for Energy Use Intensity (EUI), Daylight Autonomy (DA), Continuous Daylight Autonomy (CDA), and Useful Daylight Illuminance (UDI), respectively. The results showed that adopting different shading systems improves the distribution of lighting inside the space, increases the UDI by 27%, and decreases cooling loads by 27%. The partial effect of each shading design parameter on UDI and EUI is investigated, and then suggested values of shading design parameters are offered.

Investigation of CeO2 nanoparticles on the performance enhancement of insulating oils

Integrating nanotechnology in dielectric fluid significantly inhibits losses and boosts overall dielectric fluid performance. There has been research done on the effects of introducing various nanoparticles, such as titania, alumina, silica nanodiamonds, etc. In this paper, a novel nanoparticle, Ceria (CeO2), has been used, and its properties were examined using the FTIR (Fourier Transform Infrared) spectrum, the XRD (X-ray Diffraction) spectrum, the SEM (Scanning Electron Microscopy), and the TEM (Transmission Electron Microscopy). This paper illustrates an efficient dielectric fluid prepared by the successful dispersion of Cerium Oxide (CeO2) nanoparticles in various concentrations into four commercial oils, namely mineral oil, rapeseed oil, synthetic ester oil, and soybean oil, to enhance and improve their dielectric characteristics. The performance investigation emphasises breakdown strength enhancement and other dielectric properties of the colloidal solution comprising different nanoparticle (NP) concentrations. Various commercial oils are used as a base in nano-oil to diversify their applicability as dielectric fluids by measuring the correlation in dielectric parameters and statistically assessing their applicability with normal and Weibull distributions. The obtained experimental data sets were analyzed using the Statistics and Machine Learning Toolbox in MATLAB. The aging measurement has been done only on mineral oil, and results were matched using a predictive model of statistics and the Machine Learning Toolbox in MATLAB. Well-dispersed CeO2 NPs in the insulating oils lead to a significant increase in AC breakdown strength. The effect of ageing on the dielectric properties of nano oils yields better results than conventionally aged oil. It has been observed that the breakdown voltage is enhanced by up to 30% for mineral oil at an optimal concentration of 0.01 g/L, 9% for synthetic ester oil at 0.03 g/L, 18% for rapeseed oil at 0.02 g/L, and 19% for soybean oil at 0.03 g/L nanoparticle concentration. Following the dispersion of CeO2 nanoparticles, the dielectric constant of all insulating oils has also significantly improved.The overall experimental results are promising and show the potential of the CeO2 NPs-based nano oil as an efficient and highly performing dielectric oil for different power applications.

Groundwater levels forecasting using machine learning models: A case study of the groundwater region 10 at Karst Belt, South Africa

The crucial role which groundwater resource plays in our environment and the overall well-being of all living things can not be underestimated. Nonetheless, mismanagement of resources, over-exploitation, inadequate supply of surface water and pollution have led to severe drought and an overall drop in groundwater resources’ levels over the past decades. To address this, a groundwater flow model and several mathematical data-driven models have been developed for forecasting groundwater levels. However, there is a problem of unavailability and scarcity of the on-site input data needed by the data-driven models to forecast the groundwater level. Furthermore, as a result of the dynamics and stochastic characteristics of groundwater, there is a need for an appropriate, accurate and reliable forecasting model to solve these challenges. Over the years, the broad application of Machine Learning (ML) and Artificial Intelligence (AI) models are gaining attraction as an alternative solution for forecasting groundwater levels. Against this background, this article provides an overview of forecasting methods for predicting groundwater levels. Also, this article uses ML models such as Regressions Models, Deep Auto-Regressive models, and Nonlinear Autoregressive Neural Networks with External Input (NARX) to forecast groundwater levels using the groundwater region 10 at Karst belt in South Africa as a case study. This was done using Python Mx. Version 1.9.1., and MATLAB R2022a machine learning toolboxes. Moreover, the Coefficient of Determination (R2);, Root Mean Square Error (RMSE), Mutual Information gain, Mean Absolute Percentage Error (MAPE), Mean Squared Error (MSE), Mean Absolute Error (MAE), and the Mean Absolute Scaled Error (MASE)) models were the forecasting statistical performance metrics used to assess the predictive performance of these models. The results obtained showed that NARX and Support Vector Machine (SVM) have higher performance metrics and accuracy compared to other models used in this study.

Rapid extraction of time-dependent behaviors of endo-lysosomal structures and their cargo aided by deep learning.

The diffusional properties of subcellular structures correlate with functional properties, such as assembly of growing structures, characteristics of enclosing membranes, transient interactions with cytoskeletal elements or organelles, or capture by molecular motors. However, due to the complex, heterogeneous diffusional behavior displayed by subcellular structures, such as endo-lysosomes, it is a daunting task and major bottleneck to extract their time-dependent diffusional patterns and thus functional properties. A first step developed recently by Pinholt et al. 2021 (PNAS) to resolve this bottleneck uses a machine-learning toolbox to extract key diffusional descriptors from two-dimensional coordinates of the objects. Here we extend this work, by developing a novel deep learning modelling framework able to automatically deconvolute complex time-dependent behavior in 2D and 3D. Using this model, we correlate endosomal trafficking pathways with the diffusional properties of cargo contained within. To develop the analytical pipeline, we used 3D time-series acquired using fast live-cell lattice light sheet fluorescence microscopy of cells expressing gene-edited endo-lysosomal markers that had been infected with VSV-SARS-CoV-2 viruses tagged with other fluorescent labels. The current version of our single particle tracking end-to-end neural network model successfully ranks the key descriptive features of each trajectory point by point along the trace and takes seconds to analyze thousands of traces. It succeeded in agnostically identifying early steps of virus infection for all viral particles, including endocytosis and genome delivery into the cytosol upon virus-endosome membrane fusion.

Influence of measurement uncertainty on machine learning results demonstrated for a smart gas sensor

Abstract. Humans spend most of their lives indoors, so indoor air quality (IAQ) plays a key role in human health. Thus, human health is seriously threatened by indoor air pollution, which leads to 3.8 ×106 deaths annually, according to the World Health Organization (WHO). With the ongoing improvement in life quality, IAQ monitoring has become an important concern for researchers. However, in machine learning (ML), measurement uncertainty, which is critical in hazardous gas detection, is usually only estimated using cross-validation and is not directly addressed, and this will be the main focus of this paper. Gas concentration can be determined by using gas sensors in temperature-cycled operation (TCO) and ML on the measured logarithmic resistance of the sensor. This contribution focuses on formaldehyde as one of the most relevant carcinogenic gases indoors and on the sum of volatile organic compounds (VOCs), i.e., acetone, ethanol, formaldehyde, and toluene, measured in the data set as an indicator for IAQ. As gas concentrations are continuous quantities, regression must be used. Thus, a previously published uncertainty-aware automated ML toolbox (UA-AMLT) for classification is extended for regression by introducing an uncertainty-aware partial least squares regression (PLSR) algorithm. The uncertainty propagation of the UA-AMLT is based on the principles described in the Guide to the Expression of Uncertainty in Measurement (GUM) and its supplements. Two different use cases are considered for investigating the influence on ML results in this contribution, namely model training with raw data and with data that are manipulated by adding artificially generated white Gaussian or uniform noise to simulate increased data uncertainty, respectively. One of the benefits of this approach is to obtain a better understanding of where the overall system should be improved. This can be achieved by either improving the trained ML model or using a sensor with higher precision. Finally, an increase in robustness against random noise by training a model with noisy data is demonstrated.

Analysis of data quoted on the Day-Ahead Market of TGE S.A. using Statistics and Machine Learning Toolbox

Analysis of data quoted on the Day-Ahead Market of TGE S.A. using Statistics and Machine Learning Toolbox

Sistema de verificación de huellas dactilares basado en DWT, extracción de características de múltiples dominios y clasificador subespacial de conjuntos

This paper describes a fingerprint verification system including preprocessing, Wavelet transform, feature extraction using multiple domains, and ensemble subspace discriminant classifier. The system is implemented in MATLAB using Wavelet Toolbox, Image Processing Toolbox, and Statistics and Machine Learning Toolbox. First, the motivation and novelty, followed by the review of the previous work, are presented. Next, all steps are described in detail. Three fingerprint databases from the literature are used. The proposed method’s performance is compared with state-of-the-art techniques based on different classifiers utilizing the accuracy metric. The proposed algorithm achieves high accuracy at 97.5% for the DB3-FVC2000 subset.

Machine learning assisted interferometric structured illumination microscopy for dynamic biological imaging

Structured Illumination Microscopy, SIM, is one of the most powerful optical imaging methods available to visualize biological environments at subcellular resolution. Its limitations stem from a difficulty of imaging in multiple color channels at once, which reduces imaging speed. Furthermore, there is substantial experimental complexity in setting up SIM systems, preventing a widespread adoption. Here, we present Machine-learning Assisted, Interferometric Structured Illumination Microscopy, MAI-SIM, as an easy-to-implement method for live cell super-resolution imaging at high speed and in multiple colors. The instrument is based on an interferometer design in which illumination patterns are generated, rotated, and stepped in phase through movement of a single galvanometric mirror element. The design is robust, flexible, and works for all wavelengths. We complement the unique properties of the microscope with an open source machine-learning toolbox that permits real-time reconstructions to be performed, providing instant visualization of super-resolved images from live biological samples.

The Use of Machine Learning for Comparative Analysis of Amperometric and Chemiluminescent Methods for Determining Antioxidant Activity and Determining the Phenolic Profile of Wines

This paper presents an analysis of modern methods used to determine antioxidant activity. According to research by the World Health Organization, the deficiency of such important nutrients as antioxidants leads to a decrease in body resistance and the development of chronic diseases. When it comes to diet, the inclusion of foods with a high content of antioxidants helps to increase life expectancy. As a result of this research, the mass concentration of phenolic substances and the antioxidant activity of phenolic antioxidants in young white and red table wine materials were determined using amperometric and chemiluminescent methods in order to determine antioxidant activity. Regression equations reflecting the relationship between the indicator of antioxidant activity and the value of the mass concentration of phenolic substances in young table wine materials were derived. The conversion coefficient for determining the mass concentration of phenolic substances when using Trolox-C and gallic acid as standards was established, which was—3.75. Based on a multiple linear regression model, the total antioxidant activity of the samples (F9.5 = 19.10 and p = 0.0023) can be fairly accurately predicted with an R2 of 0.921 for the calibration data set. A neural network regression model (NNRM) was chosen for the machine-learning regression analysis of the antioxidant activity of the wine samples due to its effectiveness in predicting outcomes in various applications. The implementation was performed using the fitrnet function provided in the Statistics and Machine Learning Toolbox in MATLAB R2021b. The MSE of the calibration model was 0.056; however, the MSE for the three validation samples was much higher, at 0.272.

Uncertainty-aware automated machine learning toolbox

Abstract Measurement data can be considered complete only with an associated measurement uncertainty to express knowledge about the spread of values reasonably attributed to the measurand. Measurement uncertainty also allows to assess the comparability and the reliability of measurement results as well as to evaluate decisions based on the measurement result. Artificial Intelligence (AI) methods and especially Machine Learning (ML) are often based on measurements, but so far, uncertainty is widely neglected in this field. We propose to apply uncertainty propagation in ML to allow estimating the uncertainty of ML results and, furthermore, an optimization of ML methods to minimize this uncertainty. Here, we present an extension of a previously published automated ML toolbox (AMLT), which performs feature extraction, feature selection and classification in an automated way without any expert knowledge. To this end, we propose to apply the principles described in the “Guide to the Expression of Uncertainty in Measurement” (GUM) and its supplements to carry out uncertainty propagation for every step in the AMLT. In previous publications we have presented the uncertainty propagation for some of the feature extraction methods in the AMLT. In this contribution, we add some more elements to this concept by also including statistical moments as a feature extraction method, add uncertainty propagation to the feature selection methods and extend it to also include the classification method, linear discriminant analysis combined with Mahalanobis distance. For these methods, analytical approaches for uncertainty propagation are derived in detail, and the uncertainty propagation for the other feature extraction and selection methods are briefly revisited. Finally, the use the uncertainty-aware AMLT is demonstrated for a data set consisting of uncorrelated measurement data and associated uncertainties.

Extending machine learning beyond interatomic potentials for predicting molecular properties.

Machine learning (ML) is becoming a method of choice for modelling complex chemical processes and materials. ML provides a surrogate model trained on a reference dataset that can be used to establish a relationship between a molecular structure and its chemical properties. This Review highlights developments in the use of ML to evaluate chemical properties such as partial atomic charges, dipole moments, spin and electron densities, and chemical bonding, as well as to obtain a reduced quantum-mechanical description. We overview several modern neural network architectures, their predictive capabilities, generality and transferability, and illustrate their applicability to various chemical properties. We emphasize that learned molecular representations resemble quantum-mechanical analogues, demonstrating the ability of the models to capture the underlying physics. We also discuss how ML models can describe non-local quantum effects. Finally, we conclude by compiling a list of available ML toolboxes, summarizing the unresolved challenges and presenting an outlook for future development. The observed trends demonstrate that this field is evolving towards physics-based models augmented by ML, which is accompanied by the development of new methods and the rapid growth of user-friendly ML frameworks for chemistry.

Codebase release 3.4 for NetKet

We introduce version 3 of NetKet, the machine learning toolbox for many-body quantum physics. NetKet is built around neural quantum states and provides efficient algorithms for their evaluation and optimization. This new version is built on top of JAX, a differentiable programming and accelerated linear algebra framework for the Python programming language. The most significant new feature is the possibility to define arbitrary neural network ansätze in pure Python code using the concise notation of machine-learning frameworks, which allows for just-in-time compilation as well as the implicit generation of gradients thanks to automatic differentiation. NetKet 3 also comes with support for GPU and TPU accelerators, advanced support for discrete symmetry groups, chunking to scale up to thousands of degrees of freedom, drivers for quantum dynamics applications, and improved modularity, allowing users to use only parts of the toolbox as a foundation for their own code.

NetKet 3: Machine Learning Toolbox for Many-Body Quantum Systems

We introduce version 3 of NetKet, the machine learning toolbox for many-body quantum physics. NetKet is built around neural quantum states and provides efficient algorithms for their evaluation and optimization. This new version is built on top of JAX, a differentiable programming and accelerated linear algebra framework for the Python programming language. The most significant new feature is the possibility to define arbitrary neural network ansätze in pure Python code using the concise notation of machine-learning frameworks, which allows for just-in-time compilation as well as the implicit generation of gradients thanks to automatic differentiation. NetKet 3 also comes with support for GPU and TPU accelerators, advanced support for discrete symmetry groups, chunking to scale up to thousands of degrees of freedom, drivers for quantum dynamics applications, and improved modularity, allowing users to use only parts of the toolbox as a foundation for their own code.