Application Of Learning Algorithms Research Articles

Machine learning algorithms applied to wildfire data in California's central valley

This study focuses on using Machine Learning methods to predict wildfires within California's Central Valley. The specific areas within the Central Valley were Yosemite Valley, Sequoias, and Kings Canyon since these areas can be considered wildfire hotspots. This topic is relevant since California has seen an increase in wildfires with an increase in annual forest burned areas to +172 % from 1996 to 2021 (ABC 2024). The algorithms selected were based on previous research that conducted similar studies. From this research it is hypothesized that the best performing algorithm for predicting wildfires would be Random Forest. The novelty in this study stems from focusing on the specific areas mentioned above, which is where many wildfires have occurred throughout the years. The overall goal is to determine the best machine learning algorithm to predict wildfires in the Central Valley and take the results to improve upon wildfire prevention within these regions. The methods implemented included Decision Trees, Random Forest, Naïve Bayes, and Neural Networks. The dataset was gathered from the following satellite data which include MERRA-2 and USGS Landsat 8 along with fire history from 2012 to 2023 within these regions. Utilizing the dataset in the following two variations were a random split and a chronological split of training and testing sets. The best-performing algorithm using this dataset was Decision Trees at 550 maximum splits with an F1-Score of 0.689. The F1-Score ranges between 0 and 1 with a score of 0.7 or higher being deemed a good model to be used for predictions. The conclusion that could be determined from this result is that the randomized data has better predicting power over a chronologically split dataset. This can be seen in the confusion matrices for the chronological split dataset having zero true positive values in all the methods except for Naïve Bayes. Overall, the results show that Decision trees with a larger maximum split in the leaf nodes result in a more accurate prediction of whether a fire will occur within the given regions. The conclusion that can be made from this result is that Decision Trees can be a useful tool in predicting wildfires in California's Central Valley. The applications of this research would be the ability to use the information gained in this study to aid in optimizing resources in wildfire prevention within these areas.

Machine Learning Algorithms Applied to Predict Autism Spectrum Disorder Based on Gut Microbiome Composition.

The application of machine learning (ML) techniques stands as a reliable method for aiding in the diagnosis of complex diseases. Recent studies have related the composition of the gut microbiota to the presence of autism spectrum disorder (ASD), but until now, the results have been mostly contradictory. This work proposes using machine learning to study the gut microbiome composition and its role in the early diagnosis of ASD. We applied support vector machines (SVMs), artificial neural networks (ANNs), and random forest (RF) algorithms to classify subjects as neurotypical (NT) or having ASD, using published data on gut microbiome composition. Naive Bayes, k-nearest neighbors, ensemble learning, logistic regression, linear regression, and decision trees were also trained and validated; however, the ones presented showed the best performance and interpretability. All the ML methods were developed using the SAS Viya software platform. The microbiome's composition was determined using 16S rRNA sequencing technology. The application of ML yielded a classification accuracy as high as 90%, with a sensitivity of 96.97% and specificity reaching 85.29%. In the case of the ANN model, no errors occurred when classifying NT subjects from the first dataset, indicating a significant classification outcome compared to traditional tests and data-based approaches. This approach was repeated with two datasets, one from the USA and the other from China, resulting in similar findings. The main predictors in the obtained models differ between the analyzed datasets. The most important predictors identified from the analyzed datasets are Bacteroides, Lachnospira, Anaerobutyricum, and Ruminococcus torques. Notably, among the predictors in each model, there is the presence of bacteria that are usually considered insignificant in the microbiome's composition due to their low relative abundance. This outcome reinforces the conventional understanding of the microbiome's influence on ASD development, where an imbalance in the composition of the microbiota can lead to disrupted host-microbiota homeostasis. Considering that several previous studies focused on the most abundant genera and neglected smaller (and frequently not statistically significant) microbial communities, the impact of such communities has been poorly analyzed. The ML-based models suggest that more research should focus on these less abundant microbes. A novel hypothesis explains the contradictory results in this field and advocates for more in-depth research to be conducted on variables that may not exhibit statistical significance. The obtained results seem to contribute to an explanation of the contradictory findings regarding ASD and its relation with gut microbiota composition. While some research correlates higher ratios of Bacillota/Bacteroidota, others find the opposite. These discrepancies are closely linked to the minority organisms in the microbiome's composition, which may differ between populations but share similar metabolic functions. Therefore, the ratios of Bacillota/Bacteroidota regarding ASD may not be determinants in the manifestation of ASD.

A Machine Learning Algorithm Applied to Trichoscopy for Androgenic Alopecia Staging and Severity Assessment.

Androgenic alopecia (AGA) staging is still based on macroscopic scales, yet the introduction of trichoscopy is gradually bringing an important change, even though it remains an eye-based method. However, recently developed artificial intelligence-assisted programs can execute automated count of trichoscopic patterns. Nevertheless, to interpret data elaborated by these programs can be complex. Machine learning algorithms might represent an innovative solution. Among them, support vector machine (SVM) models are among the best methods for classification. Our aim was to develop a SVM algorithm, based on three trichoscopic patterns, able to classify AGA patients and to calculate a severity index. We retrospectively analyzed trichoscopic images from 200 AGA patients using Trichoscale Pro® software, calculating the number of vellus hair, empty follicles and single hair follicular units. Then, we elaborated a SVM model, based on these three patterns and on sex, able to classify patients as affected by mild AGA or moderate-severe AGA, and able to calculate the probability of the classification being correct, expressed as percentage (from 50% to 100%). This probability estimate is higher in patients with more AGA trichoscopic patterns and, thus, it might serve as a severity index. For training and test datasets, accuracy was 94.3% and 90.0% respectively, while the Area Under the Curve was 0.99 and 0.95 respectively. We believe our SVM model could be of great support for dermatologists in the management of AGA, especially in better assessing disease severity and, thus, in prescribing a more appropriate therapy.

Systematic Performance Evaluation of Reinforcement Learning Algorithms Applied to Wastewater Treatment Control Optimization.

Treatment of wastewater using activated sludge relies on several complex, nonlinear processes. While activated sludge systems can provide high levels of treatment, including nutrient removal, operating these systems is often challenging and energy intensive. Significant research investment has been made in recent years into improving control optimization of such systems, through both domain knowledge and, more recently, machine learning. This study leverages a novel interface between a common process modeling software and a Python reinforcement learning environment to evaluate four common reinforcement learning algorithms for their ability to minimize treatment energy use while maintaining effluent compliance within the Benchmark Simulation Model No. 1 (BSM1) simulation. Three of the algorithms tested, deep Q-learning, proximal policy optimization, and synchronous advantage actor critic, generally performed poorly over the scenarios tested in this study. In contrast, the twin delayed deep deterministic policy gradient (TD3) algorithm consistently produced a high level of control optimization while maintaining the treatment requirements. Under the best selection of state observation features, TD3 control optimization reduced aeration and pumping energy requirements by 14.3% compared to the BSM1 benchmark control, outperforming the advanced domain-based control strategy of ammonia-based aeration control, although future work is necessary to improve robustness of RL implementation.

Machine Learning Algorithms Applied to Semi-Quantitative Data of the Volatilome of Citrus and Other Nectar Honeys with the Use of HS-SPME/GC-MS Analysis, Lead to a New Index of Geographical Origin Authentication.

The scope of the current study was to monitor if semi-quantitative data of volatile compounds (volatilome) of citrus honey (ch) produced in different countries could potentially lead to a new index of citrus honey authentication using specific ratios of the identified volatile compounds in combination with machine learning algorithms. In this context, the semi-quantitative data of the volatilome of 38 citrus honey samples from Egypt, Morocco, Greece, and Spain (determined by headspace solid phase microextraction coupled to gas chromatography mass spectrometry (HS-SPME/GC-MS)) was subjected to supervised and unsupervised chemometrics. Results showed that honey samples could be classified according to the geographical origin based on specific volatile compounds. Data were further evaluated with additional nectar honey samples introduced in the multivariate statistical analysis model and the classification results were not affected. Specific volatile compounds contributed to the discrimination of citrus honey in different amounts according to geographical origin. These were lilac aldehyde D, dill ether, 2-methylbutanal, heptane, benzaldehyde, α,4-dimethyl-3-cyclohexene-1-acetaldehyde, and herboxide (isomer II). The numerical data of these volatile compounds was summed up and divided by the total semi-quantitative volatile content (Rch, Karabagias-Nayik index) of citrus honey, according to geographical origin. Egyptian citrus honey had a value of Rch = 0.35, Moroccan citrus honey had a value of Rch = 0.29, Greek citrus honey had a value of Rch = 0.04, and Spanish citrus honey had a value of Rch = 0.27, leading to a new hypothesis and a complementary index for the control of citrus honey authentication.

Retracted: Computed Tomography Image Features under Deep Learning Algorithm Applied in Staging Diagnosis of Bladder Cancer and Detection on Ceramide Glycosylation.

[This retracts the article DOI: 10.1155/2022/7979523.].

Learning from unbalanced catastrophic out-of-pocket health expenditure dataset: blending SMOTE-boosting with ensemble models

ABSTRACT This study attests to the benefits of synthetic data generation with the Synthetic Minority Oversampling Technique (SMOTE), and it incorporates this procedure with SMOTEBoosting by applying learning algorithms to model unbalanced catastrophic out-of-pocket (OOP) health expenditure dataset. Nationally representative household budget survey data were gathered from Turkish Statistical Institute for the year 2012. A total of 9987 households responded to the survey. The original dataset was highly unbalanced and a total of 0.14% of households faced catastrophic health expenses. SMOTE was used to perform balanced oversampling, and 10 artificial datasets with sizes from 10% to 100% of the majority group of original training data were generated. To predict OOP catastrophic health expenditures, the SMOTEBoosting was embedded with learning algorithms, such as C5.0, random forest (RF), naïve Bayes, and support vector machine. Study results confirm the outstanding prediction performance of the blended strategy of SMOTEBoosting with RF (area under the curve ˃ 0.85) for prediction. A variable importance plot and decision tree visualise that at least 65 years of age is the most important predictor of the catastrophic cases. The findings of this study highlight that multistrategy ensemble learning techniques are useful to model highly unbalanced datasets.

Recent Progress in Learning Algorithms Applied in Energy Management of Hybrid Vehicles: A Comprehensive Review

Recent Progress in Learning Algorithms Applied in Energy Management of Hybrid Vehicles: A Comprehensive Review

Parametric Validation of the Reservoir Computing–Based Machine Learning Algorithm Applied to Lorenz System Reconstructed Dynamics

A detailed parametric analysis is presented, where the recent method based on the reservoir computing paradigm, including its statistical robustness, is studied. It is observed that the prediction capabilities of the reservoir computing approach strongly depend on the random initialization of both the input and the reservoir layers. Special emphasis is put on finding the region in the hyperparameter space where the ensemble-averaged training and generalization errors together with their variance are minimized. The statistical analysis presented here is based on the projection on proper elements method.

A Needs Learning Algorithm Applied to Stable Gait Generation of Quadruped Robot

Based on Maslow’s hierarchy of needs theory, we have proposed a novel machine learning algorithm that combines factors of the environment and its own needs to make decisions for different states of an agent. This means it can be applied to the gait generation of a quadruped robot, which needs to make demand decisions. To evaluate the design, we created an experimental task in order to compare the needs learning algorithm with a reinforcement learning algorithm, which was also derived from psychological motivation theory. It was found that the needs learning algorithm outperformed the reinforcement learning in tasks that involved making decisions between different levels of needs. Finally, we applied the needs learning algorithm to the problem of stable gait generation of quadruped robot, and it had achieved good results in simulation and real robot.

Views of Deep Learning Algorithm Applied to Computer Vision Knowledge Discovery

Views of Deep Learning Algorithm Applied to Computer Vision Knowledge Discovery

The use of artificial intelligence methods in the analysis of the results of vehicle braking deceleration tests in the diagnostics of the braking system of a motor vehicle

The article presents the concept of using an artificial neural network to approximate the parameters describing the vehicle braking process, from the point of view of the application of this method in the diagnostics of the braking system. The artificial neural network of non-linear autoregression was used to approximate the dependence of the braking deceleration and the pressure in the braking system. The effectiveness of the neural network was checked depending on the number of neurons in its hidden layer and on the applied learning algorithm. The operation of the neural network was verified based on the actual braking processes of the Skoda Octavia, carried out with different dynamics, with different car weights and different tire inflation pressures. After verifying the neural network, it was used to approximate the braking deceleration values for the pressure values exceeding those present in the input data set. This action allows the analysis of the possibility of the vehicle obtaining a braking deceleration, which qualifies its braking system as efficient. Two concepts of using a neural network to solve this problem were analyzed. Conclusions related to the validity of the development of the discussed methods were drawn.

SafeDriving: An Effective Abnormal Driving Behavior Detection System Based on EMG Signals

To improve safety in public transportation, a major issue is how to avoid traffic accidents. To this end, a recent report has demonstrated that more than 90% of accidents in the United States were due to drivers’ abnormal behaviors. Relevant to this observation, many recent studies have proposed to use different sensors to monitor drivers’ behaviors and apply learning algorithms to detect abnormal behaviors. Nevertheless, most existing systems are expensive and inconvenient to be deployed or significantly affected by the environment. In this article, we propose and develop a novel and effective solution, namely, SafeDriving, that collects signals from electromyography (EMG) sensors and then utilizes an effective deep-learning model to detect abnormal behaviors in real time. Specifically, we first utilize a wearable EMG sensor that can be attached to a driver’s forearm to collect a large amount of sensing data from human drivers, for which we define five typical abnormal driving behaviors (i.e., fetching forward, picking up, turning the steering wheel sharply, turning back, and touching sunroof) and label each sample accordingly. Next, using the labeled data, we design and train multiple state-of-the-art classifiers to improve the performance of SafeDriving, e.g., convolutional neural network (CNN), long short-term memory (LSTM), and gated recurrent unit (GRU). The extensive experiments demonstrate that GRU can lead to the best performance with an average accuracy of 93.94%. Based on this observation, we further investigate other important factors, such as the binding area of the sensor, the tightness of binding, the duration of the sample, etc. The proposed SafeDriving system provides an effective approach to reliably assess drivers’ driving behaviors with affordable commodity sensors and be further used in public safety.

Computed Tomography Image Features under Deep Learning Algorithm Applied in Staging Diagnosis of Bladder Cancer and Detection on Ceramide Glycosylation.

The research is aimed at investigating computed tomography (CT) image based on deep learning algorithm and the application value of ceramide glycosylation in diagnosing bladder cancer. The images of ordinary CT detection were improved. In this study, 60 bladder cancer patients were selected and performed with ordinary CT detection, and the detection results were processed by CT based on deep learning algorithms and compared with pathological diagnosis. In addition, Western Blot technology was used to detect the expression of glucose ceramide synthase (GCS) in the cell membrane of tumor tissues and normal tissues of bladder. The comparison results found that, in simple CT clinical staging, the coincidence rates of T1 stage, T2a stage, T2b stage, T3 stage, and T4 stage were 28.56%, 62.51%, 78.94%, 84.61%, and 74.99%, respectively; and the total coincidence rate of CT clinical staging was 63.32%, which was greatly different from the clinical staging of pathological diagnosis (P < 0.05). In the clinical staging of algorithm-based CT test results, the coincidence rates of T1 stage and T2a stage were 50.01% and 91.65%, respectively; and those of T2b stage, T3 stage, and T4 stage were 100.00%; and the total coincidence rate was 96.69%, which was not obviously different from the clinical staging of pathological diagnosis (P > 0.05). Therefore, it could be concluded that the algorithm-based CT detection results were more accurate, and the use of CT scans based on deep learning algorithms in the preoperative staging and clinical treatment of bladder cancer showed reliable guiding significance and clinical value. In addition, it was found that the expression level of GCS in normal bladder tissues was much lower than that in bladder cancer tissues. This indicated that the changes in GCS were closely related to the development and prognosis of bladder cancer. Therefore, it was believed that GCS may be an effective target for the treatment of bladder cancer in the future, and further research was needed for specific conditions.

Assessing the Reidentification Risks Posed by Deep Learning Algorithms Applied to ECG Data

ECG (Electrocardiogram) data analysis is one of the most widely used and important tools in cardiology diagnostics. In recent years the development of advanced deep learning techniques and GPU hardware have made it possible to train neural network models that attain exceptionally high levels of accuracy in complex tasks such as heart disease diagnoses and treatments. We investigate the use of ECGs as biometrics in human identification systems by implementing state-of-the-art deep learning models. We train convolutional neural network models on approximately 81k patients from the US, Germany and China. Currently, this is the largest research project on ECG identification. Our models achieved an overall accuracy of 95.69%. Furthermore, we assessed the accuracy of our ECG identification model for distinct groups of patients with particular heart conditions and combinations of such conditions. For example, we observed that the identification accuracy was the highest (99.7%) for patients with both ST changes and supraventricular tachycardia. We also found that the identification rate was the lowest for patients diagnosed with both atrial fibrillation and complete right bundle branch block (49%). We discuss the implications of these findings regarding the reidentification risks of patients based on ECG data and how seemingly anonymized ECG datasets can cause privacy concerns for the patients.

Health Data-Driven Machine Learning Algorithms Applied to Risk Indicators Assessment for Chronic Kidney Disease.

PurposeAs global aging progresses, the health management of chronic diseases has become an important issue of concern to governments. Influenced by the aging of its population and improvements in the medical system and healthcare in general, Taiwan’s population of patients with chronic kidney disease (CKD) has tended to grow year by year, including the incidence of high-risk cases that pose major health hazards to the elderly and middle-aged populations.MethodsThis study analyzed the annual health screening data for 65,394 people from 2010 to 2015 sourced from the MJ Group – a major health screening center in Taiwan – including data for 18 risk indicators. We used five prediction model analysis methods, namely, logistic regression (LR) analysis, C5.0 decision tree (C5.0) analysis, stochastic gradient boosting (SGB) analysis, multivariate adaptive regression splines (MARS), and eXtreme gradient boosting (XGboost), with estimated glomerular filtration rate (e-GFR) data to determine G3a, G3b & G4 stage CKD risk factors.ResultsThe LR analysis (AUC=0.848), SGB analysis (AUC=0.855), and XGboost (AUC=0.858) generated similar classification performance levels and all outperformed the C5.0 and MARS methods. The study results showed that in terms of CKD risk factors, blood urea nitrogen (BUN) and uric acid (UA) were identified as the first and second most important indicators in the models of all five analysis methods, and they were also clinically recognized as the major risk factors. The results for systolic blood pressure (SBP), SGPT, SGOT, and LDL were similar to those of a related study. Interestingly, however, socioeconomic status-related education was found to be the third important indicator in all three of the better performing analysis methods, indicating that it is more important than the other risk indicators of this study, which had different levels of importance according to the different methods.ConclusionThe five prediction model methods can provide high and similar classification performance in this study. Based on the results of this study, it is recommended that education as the socioeconomic status should be an important factor for CKD, as high educational level showed a negative and highly significant correlation with CKD. The findings of this study should also be of value for further discussions and follow-up research.

Machine Learning MOSA Monitoring System

This research paper deals with the problem of Metal-Oxide Surge Arrester (MOSA) condition monitoring and a new methodology in surge arrester monitoring and diagnostics is presented. A machine learning algorithm (back propagation regression) is used to estimate the non-linearity coefficient of the surge arrester, based on operating voltage and leakage current of the arrester. Using a simulated system, this research investigates the possibility of application and efficiency of machine learning. It is shown that the applied learning algorithm results are competitive with the model results parameters calculated as R2 = 0.999 and mean absolute real error computed as 0.005 which has shown that the proposed model can be used for MOSA monitoring and diagnostic purposes.

PolyG2G: A Novel Machine Learning Algorithm Applied to the Generative Design of Polymer Dielectrics

Polymers, due to advantages such as low-cost processing, chemical stability, low density, and tunable design, have emerged as a powerhouse class of materials for a wide range of applications, including dielectrics. However, in certain applications, the performance of dielectrics is limited by insufficient electric breakdown strength. Using this real-world application as a technology driver, we describe a novel artificial intelligence (AI)-based approach for the design of polymers. We call this approach polyG2G. The key concept underlying polyG2G is graph-to-graph translation. Graph-to-graph translation solves the inverse problem. First, the subtle chemical differences between high- and low-performing polymers are learned. Then, the learned differences are applied to known polymers, yielding large libraries of novel, high-performing, hypothetical polymers. Our approach, with respect to a host of presently adopted design methods, exhibits a favorable trade-off between generation of chemically valid materials and available chemical search space. polyG2G finds thousands of potentially high-value targets (in terms of glass-transition temperature, band gap, and electron injection barrier) from an otherwise intractable search space. Density functional theory simulations of band gap and electron injection barrier confirm that a large fraction of the polymers designed by polyG2G are indeed of high value. Finally, we find that polyG2G is able to learn established structure–property relationships.

Manifold Learning Algorithms Applied to Structural Damage Classification

A comparative study of four manifold learning algorithms was carried out to perform the dimensionality reduction process within a proposed methodology for damage classification in structural health monitoring (SHM). Isomap, locally linear embedding (LLE), stochastic proximity embedding (SPE), and laplacian eigenmaps were used as manifold learning algorithms. The methodology included several stages that comprised: data normalization, dimensionality reduction, classification through K-Nearest Neighbors (KNN) machine learning model and finally holdout cross-validation with 25% of data for training and the remaining 75% of data for testing. Results evaluated in an experimental setup showed that the best classification accuracy was 100% when the methodology uses isomap algorithm with a hyperparameter k of 170 and 8 dimensions as a feature vector at the input to the KNN classification machine.

Machine Learning Algorithms Applied to Identify Microbial Species by Their Motility.

(1) Background: Future missions to potentially habitable places in the Solar System require biochemistry-independent methods for detecting potential alien life forms. The technology was not advanced enough for onboard machine analysis of microscopic observations to be performed in past missions, but recent increases in computational power make the use of automated in-situ analyses feasible. (2) Methods: Here, we present a semi-automated experimental setup, capable of distinguishing the movement of abiotic particles due to Brownian motion from the motility behavior of the bacteria Pseudoalteromonas haloplanktis, Planococcus halocryophilus, Bacillus subtilis, and Escherichia coli. Supervised machine learning algorithms were also used to specifically identify these species based on their characteristic motility behavior. (3) Results: While we were able to distinguish microbial motility from the abiotic movements due to Brownian motion with an accuracy exceeding 99%, the accuracy of the automated identification rates for the selected species does not exceed 82%. (4) Conclusions: Motility is an excellent biosignature, which can be used as a tool for upcoming life-detection missions. This study serves as the basis for the further development of a microscopic life recognition system for upcoming missions to Mars or the ocean worlds of the outer Solar System.