Random Decision Forest Research Articles

Machine learning models to predict 30-day mortality for critical patients with myocardial infarction: a retrospective analysis from MIMIC-IV database.

The identification of efficient predictors for short-term mortality among patients with myocardial infarction (MI) in coronary care units (CCU) remains a challenge. This study seeks to investigate the potential of machine learning (ML) to improve risk prediction and develop a predictive model specifically tailored for 30-day mortality in critical MI patients. This study focused on MI patients extracted from the Medical Information Mart for Intensive Care-IV database. The patient cohort was randomly stratified into derivation (n = 1,389, 70%) and validation (n = 595, 30%) groups. Independent risk factors were identified through eXtreme Gradient Boosting (XGBoost) and random decision forest (RDF) methodologies. Subsequently, multivariate logistic regression analysis was employed to construct predictive models. The discrimination, calibration and clinical utility were assessed utilizing metrics such as receiver operating characteristic (ROC) curve, calibration plot and decision curve analysis (DCA). A total of 1,984 patients were identified (mean [SD] age, 69.4 [13.0] years; 659 [33.2%] female). The predictive performance of the XGBoost and RDF-based models demonstrated similar efficacy. Subsequently, a 30-day mortality prediction algorithm was developed using the same selected variables, and a regression model was visually represented through a nomogram. In the validation group, the nomogram (Area Under the Curve [AUC]: 0.835, 95% Confidence Interval [CI]: [0.774-0.897]) exhibited superior discriminative capability for 30-day mortality compared to the Sequential Organ Failure Assessment (SOFA) score [AUC: 0.735, 95% CI: (0.662-0.809)]. The nomogram (Accuracy: 0.914) and the SOFA score (Accuracy: 0.913) demonstrated satisfactory calibration. DCA indicated that the nomogram outperformed the SOFA score, providing a net benefit in predicting mortality. The ML-based predictive model demonstrated significant efficacy in forecasting 30-day mortality among MI patients admitted to the CCU. The prognostic factors identified were age, blood urea nitrogen, heart rate, pulse oximetry-derived oxygen saturation, bicarbonate, and metoprolol use. This model serves as a valuable decision-making tool for clinicians.

Liver Disease Detection using Machine Learning

Abstract: Liver disease is a serious worldwide health issue, and prompt diagnosis and treatment are essential for successful outcomes. Traditional diagnostic techniques, however, may be expensive and time-consuming and can require intrusive procedures. In this work, we suggest a machine learning-based method for liver disease identification that makes use of the Support Vector Machine (SVM) and Random Forest Decision Tree algorithms. Our approach uses a large dataset with pertinent clinical characteristics including biochemical signs and patient demographics to categorize people into liver disease-positive or - negative groups. Furthermore, we incorporate an intuitive UI with Streamlit, improving accessibility and usability for end users and healthcare providers alike. By means of meticulous testing and assessment, we exhibit the efficiency and functionality of our suggested framework. In terms of precision, specificity, and sensitivity, our model performs admirably, demonstrating its promise as a trustworthy instrument for the identification of liver illness. Moreover, the incorporation of Streamlit improves the system's practicality and usability, making it easier to install and use in actual healthcare environments.

Evaluation of Machine Learning Classification Models for False-Positive Reduction in Prostate Cancer Detection Using MRI Data.

In this work, several machine learning (ML) algorithms, both classical ML and modern deep learning, were investigated for their ability to improve the performance of a pipeline for the segmentation and classification of prostate lesions using MRI data. The algorithms were used to perform a binary classification of benign and malignant tissue visible in MRI sequences. The model choices include support vector machines (SVMs), random decision forests (RDFs), and multi-layer perceptrons (MLPs), along with radiomic features that are reduced by applying PCA or mRMR feature selection. Modern CNN-based architectures, such as ConvNeXt, ConvNet, and ResNet, were also evaluated in various setups, including transfer learning. To optimize the performance, different approaches were compared and applied to whole images, as well as gland, peripheral zone (PZ), and lesion segmentations. The contribution of this study is an investigation of several ML approaches regarding their performance in prostate cancer (PCa) diagnosis algorithms. This work delivers insights into the applicability of different approaches for this context based on an exhaustive examination. The outcome is a recommendation or preference for which machine learning model or family of models is best suited to optimize an existing pipeline when the model is applied as an upstream filter.

Using near misses, artificial intelligence, and machine learning to predict maritime incidents: A U.S. Coast Guard case study.

Two recent trends made this project possible: (1) The recognition that near misses can be predictors of future negative events and (2) enhanced artificial intelligence (AI) and machine learning (ML) tools that make data analytics accessible for many organizations. Increasingly, organizations are learning from prior incidents to improve safety and reduce accidents. The U.S. Coast Guard (USCG) uses a reporting system called the Marine Information for Safety and Law Enforcement (MISLE) database. Because many of the incidents that appear in this database are minor ones, this project initially focused on determining if near misses in MISLE could be predictors of future accidents. The analysis showed that recent near-miss counts are useful for predicting future serious casualties at the waterway level. Using this finding, a predictive AI/ML model was built for each waterway type by vessel combination. Random forest decision tree AI/ML models were used to identify waterways at significant accident risk. An R-based predictive model was designed to be run monthly, using data from prior months to make future predictions. The prediction models were trained on data from 2007 to 2022 and tested on 10 months of data from 2022, where prior months were added to test the next month. The overall accuracy of the predictions was 92%-99.9%, depending on model characteristics. The predictions of the models were considered accurate enough to be potentially useful in future prevention efforts for the USCG and may be generalizable to other industries that have near-miss data and a desire to identify and manage risks.

Comparative analysis of gene regulatory networks identifies conserved regulators in seed plants

Abstract Gene regulatory networks (GRNs) based on transcription factors (TFs) control development and environmental responses. In this study, GRNs were inferred computationally using random forest decision tree-based regression. Networks were constructed for the grasses barley, maize, wheat, brachypodium, sorghum and rice. When compared with Arabidopsis thaliana and alga networks, they show substantial conservation. The degree of conservation depends on phylogenetic closeness. The processes are conserved between all species include basic cellular functions while the processes conserved only in the grasses are annotated with more specific gene ontology terms. In the three species with a carbon concentration mechanism, photorespiration is partially dissociated from photosynthetic regulation. In contrast, in the C4 species, the regulation of C4 genes is associated with photosynthetic regulation. The comparative analyses reveal conserved TFs that control photosynthesis in seed plants but not in the alga. The analyses resulted in a pipeline for the general transfer of information between the small dicot A. thaliana and the commercially relevant grasses.

The effect of fine particle peening on aviation gear performance, Part I: Surface integrity

AbstractAs a fundamental component of aviation equipment, the gear has an essential effect on the service performance of the whole machine. Fine particle peening (FPP) technology is a good choice for the precision requirements of aviation gears. Therefore, a FPP model of aviation gears was constructed based on the discrete element‐finite element coupling method (DEM‐FEM) to simulate the evolution of surface roughness and residual compressive stress, and its results were verified by experiments. The effects of FPP intensity, coverage rate, and particle diameter on surface integrity were investigated. In addition, the importance degree of the intensity, coverage rate, and particle diameter of FPP on surface integrity was carried out by the random forest decision method. It demonstrates that the surface roughness Sa was mainly influenced by the FPP intensity and particle diameter. And the distribution state of the residual compressive stress was mainly affected by the particle diameter.

Do Acoustic Characteristics of Dysarthria in People With Parkinson's Disease Differ Across Languages?

Cross-language studies suggest more similarities than differences in how dysarthria affects the speech of people with Parkinson's disease (PwPD) who speak different languages. In this study, we aimed to identify the relative contribution of acoustic variables to distinguish PwPD from controls who spoke varieties of two Romance languages, French and Portuguese. This bi-national, cross-sectional, and case-controlled study included 129 PwPD and 124 healthy controls who spoke French or Portuguese. All participants underwent the same clinical examinations, voice/speech recordings, and self-assessment questionnaires. PwPD were evaluated off and on optimal medication. Inferential analyses included Disease (controls vs. PwPD) and Language (French vs. Portuguese) as factors, and random decision forest algorithms identified relevant acoustic variables able to distinguish participants: (a) by language (French vs. Portuguese) and (b) by clinical status (PwPD on and off medication vs. controls). French-speaking and Portuguese-speaking individuals were distinguished from each other with over 90% accuracy by five acoustic variables (the mean fundamental frequency and the shimmer of the sustained vowel /a/ production, the oral diadochokinesis performance index, the relative sound level pressure and the relative sound pressure level standard deviation of the text reading). A distinct set of parameters discriminated between controls and PwPD: for men, maximum phonation time and the oral diadochokinesis speech proportion were the most significant variables; for women, variables calculated from the oral diadochokinesis were the most discriminative. Acoustic variables related to phonation and voice quality distinguished between speakers of the two languages. Variables related to pneumophonic coordination and articulation rate were the more effective in distinguishing PwPD from controls. Thus, our research findings support that respiration and diadochokinesis tasks appear to be the most appropriate to pinpoint signs of dysarthria, which are largely homogeneous and language-universal. In contrast, identifying language-specific variables with the speech tasks and acoustic variables studied was less conclusive.

Machine Learning Methods and Visual Observations to Categorize Behavior of Grazing Cattle Using Accelerometer Signals.

Accelerometers worn by animals produce distinct behavioral signatures, which can be classified accurately using machine learning methods such as random forest decision trees. The objective of this study was to identify accelerometer signal separation among parsimonious behaviors. We achieved this objective by (1) describing functional differences in accelerometer signals among discrete behaviors, (2) identifying the optimal window size for signal pre-processing, and (3) demonstrating the number of observations required to achieve the desired level of model accuracy,. Crossbred steers (Bos taurus indicus; n = 10) were fitted with GPS collars containing a video camera and tri-axial accelerometers (read-rate = 40 Hz). Distinct behaviors from accelerometer signals, particularly for grazing, were apparent because of the head-down posture. Increasing the smoothing window size to 10 s improved classification accuracy (p < 0.05), but reducing the number of observations below 50% resulted in a decrease in accuracy for all behaviors (p < 0.05). In-pasture observation increased accuracy and precision (0.05 and 0.08 percent, respectively) compared with animal-borne collar video observations.

Enhance Road Detection Data Processing of LiDAR Point Clouds to Specifically Identify Unmarked Gravel Rural Roads

Abstract Gravel roads lack standardized features such as curbs or painted lines, presenting detection challenges to autonomous vehicles. Global positioning service (GPS) and high resolution maps may not be reliable for navigation of gravel roads, as some roads may only be width of the vehicle and GPS may not be accurate enough. Normal distribution transform (NDT) LiDAR scan matching may be insufficient for navigating on gravel roads as there may not be enough geometrically distinct features for reliable scan matching. This paper examined a method of classifying scanning LiDAR spatial and remission data features for explicit detection of unmarked gravel road surfaces. Exploration of terrain classification using high resolution scanning LiDAR data of specific road surfaces may allow for predicting gravel road boundary locations potentially enabling confident autonomous operations on gravel roads. The principal outcome of this work was a method for gravel road terrain detection using LiDAR data for the purpose of predicting potential road boundary locations. Random decision forests were trained using scanning LiDAR data terrain classification to detect unmarked gravel and asphalt surfaces. It was found that a true-positive accuracy for gravel and asphalt surfaces was 75% and 87%, respectively, at an estimated rate of 13 ms per 360 deg scan. Overlapping results between manually projected and actual road surface areas resulted in 93% intersecting gravel road detection accuracy. Automated post-process examination of classification results yielded an true-positive gravel road detection rate of 72%.

Islanding Detection and Power Quality Diagnosis of Wind Power Integrated Microgrid with Reduced Feature Trained Novel Optimized Random Decision Forest

Distributed generations (DGs) have been increasingly addressing the ongoing power deficit in the electricity market. However, a significant concern in DG-integrated microgrids is the detection of accidental islanding. To tackle this issue, this article proposes a cost-friendly, novel data-driven passive islanding detection scheme named EEMD-HOBRC, combining noise-assisted ensemble empirical mode decomposition (EEMD) and a hybrid optimization-based random forest classifier (HOBRFC). The detection scheme employs a diverse set of features extracted from both raw and EEMD decomposed signals. Essential features are selected using the binary grey wolf optimizer (BGWO) to reduce computational burden. To further improve classification accuracy, the parameters of the random forest classifier are optimized through a hybrid particle swarm and reformed grey wolf optimization (PSRGWO) technique with Cohen’s kappa index as the cost function. The proposed technique is rigorously validated in two different multi-DG environments, encompassing islanding and various nonislanding events. The results demonstrate the effectiveness of the approach in terms of enhanced accuracy, detection time, and performance under both noisy and noise-free conditions. The accuracy of detection under ideal and high noise scenarios is found to be 99.88% and 99.2%, respectively, with maximum detection time of 34.27 ms. Comparative analysis with other algorithms also supports the superiority of the proposed technique. Finally, the method is successfully applied to shrink the nondetection zone (NDZ) with minimal power mismatch, further enhancing its utility in practical applications.

A ROBUST CYBER SECURITY THREAT DETECTION MODEL USING ARTIFICIAL INTELLIGENCE TECHNOLOGY

The difficulty of ensuring cyber-security is steadily growing as a result of the alarming development in computer connectivity and the sizeable number of applications associated to computers in recent years. The system also requires robust defines against the growing number of cyber threats. As a result, a possible role for cyber-security might be performed by developing intrusion detection systems (ids) to detect inconsistencies and threats in computer networks. An effective data-driven intrusion detection system has been created with the use of artificial intelligence, particularly machine learning techniques. This research proposes a novel twin support vector machine (tsvm) based security model which first considers the security features ranking according to their relevance before developing an ids model based on the significant features that have been selected. By lowering the feature dimensions, this approach not only improves predictive performance for unidentified tests but also lowers the model's computational expense. Trials are conducted using four common ml techniques to compare the results to those of the current approaches (decision tree, random decision forest, random tree, and artificial neural network). The experimental findings of this study confirm that the suggested methods may be used as learning-based models for network intrusion detection and demonstrate that, when used in the real world, they outperform conventional ml techniques.

Predicting psycho-somatic disorders in online activity using multi-layer perceptron

&lt;p&gt;Internet services such as social media, blogs, and websites make it possible for people to acquire knowledge instantly. Due to these websites, it is now considerably easier to communicate information. As a result, individuals increasingly devote a higher amount of time to social networking programmes. This study provides estimates about the potential future ramifications of how individuals will utilise social networks. This work presents an accurate and applicable model for forecasting undesirable consequences. The model is of sufficient quality to be useful. This has been the case throughout. Using the model that has been proposed, significant properties are identified from datasets. After recovering the properties, they are categorised using the complicated computational method of multi-layer perceptron-based (MLP) artificial neural networks (ANN). 70% of this data was utilised during the training phase of the machine learning algorithm, while the remaining 30% was utilised during the validation phase of model construction. The proposed model's results were compared to those of more standard machine learning techniques. The approach utilises social networks to predict the issue. The simulation results indicate that the suggested model generates more precise predictions than the support vector machine, logistic regression, and random forest decision tree classifier techniques combined.&lt;/p&gt;

Deep-GD: Deep Learning based Automatic Garment Defect Detection and Type Classification

Garment defect detection has been successfully implemented in the quality quick response system for textile manufacturing automation. Defects in the production of textiles waste a lot of resources and reduce the quality of the finished goods. It is challenging to detect garment defects automatically because of the complexity of images and variety of patterns in textiles. This study presented a novel deep learning-based Garment defect detection framework named as Deep-GD model for sequentially identifying image defects in patterned garments and classify the defect types. Initially, the images are gathered from the HKBU database and bilateral filters are used in the pre-processing of images to remove distortions. A squeeze-and-excitation network (SE-net) combined with Random Decision Forest Classifier with Bayesian Optimization (RDF-BO) algorithm is used to detect and classify garment defects. By analyzing the differences among the original and reconstruction images, the unsupervised technique trains to rebuild the fabric pattern. The SE-net is used to identify certain fabric flaws in the garments of the pre-processed images. Then, the defects-related garments are processed using RDF-BO algorithm for classifying the garment defect from these regions. Finally, the proposed Deep-GD model is used for classifying defected fabric into 12 classes such as Defect free, soil stain, oil stain, Double end, Snarls, Miss, Horizontal stripes, Lumpy, Dye spot, Fall out, Hairiness, tiny hole. The proposed Deep-GD model achieves the overall classification accuracy of 97.16%, which is comparatively superior to the existing techniques.

Artificial Intelligence-Based Prediction of Contrast Medium Doses for Computed Tomography Angiography Using Optimized Clinical Parameter Sets.

In this paper, an artificial intelligence-based algorithm for predicting the optimal contrast medium dose for computed tomography (CT) angiography of the aorta is presented and evaluated in a clinical study. The prediction of the contrast dose reduction is modelled as a classification problem using the image contrast as the main feature. This classification is performed by random decision forests (RDF) and k-nearest-neighbor methods (KNN). For the selection of optimal parameter subsets all possible combinations of the 22 clinical parameters (age, blood pressure, etc.) are considered using the classification accuracy and precision of the KNN classifier and RDF as quality criteria. Subsequently, the results of the evaluation were optimized by means of feature transformation using regression neural networks (RNN). These were used for a direct classification based on regressed Hounsfield units as well as preprocessing for a subsequent KNN classification. For feature selection, an RDF model achieved the highest accuracy of 84.42% and a KNN model achieved the best precision of 86.21%. The most important parameters include age, height, and hemoglobin. The feature transformation using an RNN considerably exceeded these values with an accuracy of 90.00% and a precision of 97.62% using all 22 parameters as input. However, also the feasibility of the parameter sets in routine clinical practice has to be considered, because some of the 22 parameters are not measured in routine clinical practice and additional measurement time of 15 to 20 minutes per patient is needed. Using the standard feature set available in clinical routine the best accuracy of 86.67% and precision of 93.18% was achieved by the RNN. We developed a reliable hybrid system that helps radiologists determine the optimal contrast dose for CT angiography based on patient-specific parameters.

On the nature of apparent transient sources on the National Geographic Society–Palomar Observatory Sky Survey glass copy plates

Abstract We examine critically recent claims for the presence of above-atmosphere optical transients in publicly available digitized scans of Schmidt telescope photographic plate material derived from the National Geographic Society–Palomar Observatory Sky Survey. We employ the publicly available SuperCOSMOS Sky Survey catalogues to examine statistically the morphology of the sources. We develop a simple, objective, and automated image classification scheme based on a random forest decision tree classifier. We find that the putative transients are likely to be spurious artefacts of the photographic emulsion. We suggest a possible cause of the appearance of these images as resulting from the copying procedure employed to disseminate glass copy survey atlas sets in the era before large-scale digitization programmes.

Developing the health effect assessment of landscape (HEAL) Tool: Assessing the health effects of community greenspace morphology design on non-communicable diseases

Neighborhood greenspace benefits health, yet few tools are available for estimating the health consequences of community greenspace design alternatives, especially prior to the implementation of a landscape design plan. Herein we present a machine learning based tool for predicting the prevalence of non-communicable diseases based on landscape design maps at the community scale. We achieve this based on data collected in five major metropolises in the United States. By using high-resolution satellite imagery and remote sensing technologies, landscape spatial design characteristics were extracted through spatial pattern analysis, allowing the measurement of verdancy, fragmentation, connectedness, aggregation, and shape of greenspaces. We established a model to estimate the prevalence of poor mental health, coronary heart disease, stroke, diabetes, chronic obstructive pulmonary disease, and physical inactivity at the census tract level by adopting a combination of random forest decision tree algorithm and spatial Gaussian process models. Model accuracy was found to be significantly higher than ordinary regression models. The model accounted for very high levels of variance in the specified morbidities; viz., poor mental health (97%), heart disease (93%), stroke (93%), diabetes (95%), COPD (94%), and physical inactivity (98%). This tool is implemented using a freely available programming language (R), and we offer the model accessible to the public. This tool enables urban planners and landscape designers to assess and compare the health effects of different greenspace design plans prior to implementation, thus providing policymakers and designers with evidence-rich alternatives during the health-promoting greenspace planning process.

An efficient Cucconi based feature extraction with random decision forest classification for improved sentiment analysis

Sentiment analysis is a form of opinion mining technique that identifies the polarity of extracted opinions. Nowadays, opinion mining has become an important research area in recent decades to identify the polarity of the statements. Various research works have been carried out on sentiment analysis. However, the existing sentimental analysis techniques, such as time and space complexity, still have considerable limitations. To deal with these issues, this paper proposed the Cucconi Feature Extracted Random Decision Forest Classification (CFDFC) Approach. The main objective of the CFDFC approach is to provide effective sentiment analysis with improved accuracy and reduced time complexity. The proposed CFDFC framework comprises pre-processing, feature extraction, and classification. The pre-processing step eliminates stop words and stem words from user reviews. After the pre-processing step, the feature extraction process is carried out to minimize the dimensionality and time consumption for opinion classification. Cucconi's projective feature extraction process is used in this work to reduce dimensionality. Finally, the classification process is formulated using a random decision forest classifier. The random decision forest classifier uses the ID3 DT (decision tree) as a weak learner to classify the review statements. The performance evaluation of the proposed approach is carried out using performance metrics such as accuracy, error rates, recall values, and time and space complexities concerning the number of review statements gathered from the dataset. The results show that the proposed CFDFC model achieves remarkable accuracy, recall, and minimal time complexity compared to existing methods.

Forecasting model for qualitative prediction of the results of patient-specific quality assurance based on planning and complexity metrics and their interrelations. Pilot study.

The purpose was to analyse the interrelations between planning and complexity metrics and gamma passing rates (GPRs) obtained from VMAT treatments and build the forecasting models for qualitative prediction (QD) of GPRs results. 802 treatment arcs from the plans prepared for the head and neck, thorax, abdomen, and pelvic cancers were analysed. The plans were verified by portal dosimetry and analysed twice using the gamma method with 3%|2mm and 2%|2mm acceptance criteria. The tolerance limit of GPR was 95%. Red, yellow, and green QDs were established for GPR examination. The interrelations were examined, as well as the analysis of effective differentiation of QD. Three models for QD forecasting based on discriminant analysis (DA), random decision forest (RDF) methods, and the hybrid model (HM) were built and evaluated. Most of the interrelations were small or moderate. The exception is correlations of the join function with the average number of monitor units per control point (R = 0.893) and the beam aperture with planning target volume (R = 0.897). While many metrics allow for the effective separation of the QDs from each other, the study shows that predicting the values of the QD is possible only through multi-component forecasting models, of which the HM is the most accurate (0.894). Of the three models explored in this study, the HM, which uses DA methods to predict red QD and RDF methods to predict green and yellow QDs, is the most promising one.

Incorporating landscape ecological approach in machine learning classification for agricultural land-use mapping based on a single date imagery

Land-use maps containing crop rotational information are very important in land management and physical planning. Such maps are usually generated using multitemporal data. Although recent technology allows analysts to process multitemporal information effectively, the use of single date imagery for such purpose is more efficient. This study aimed to map detailed agricultural land-use with crop rotational information based on a single date Landsat 8 imagery and SRTM-derived terrain attributes. A landscape ecological approach assuming the influence of terrain characteristics on the existence of crop and land-use types was implemented in multisource classification using random decision forest (RDF) machine learning algorithm. The use of seven optical bands and five terrain attributes could provide a land-use map at 88.03% accuracy, compared to seven optical bands only that generate 82.45% accuracy. These results are also better than those of maximum likelihood. The most influential variables in the achieved accuracy are elevation and thermal band.

Investigating artificial intelligence models for predicting joint pain from serum biochemistry.

The study used machine learning models to predict the clinical outcome with various attributes or when the models chose features based on their algorithms. Patients who presented to an orthopedic outpatient department with joint swelling or myalgia were included in the study. A proforma collected clinical information on age, gender, uric acid, C-reactive protein, and complete blood count/liver function test/renal function test parameters. Machine learning decision models (Random Forest and Gradient Boosted) were evaluated with the selected features/attributes. To categorize input data into outputs of indications of joint discomfort, multilayer perceptron and radial basis function-neural networks were used. The random forest decision model outperformed with 97% accuracy and minimum errors to anticipate joint pain from input attributes. For predicted classifications, the multilayer perceptron fared better with an accuracy of 98% as compared to the radial basis function. Multilayer perceptron achieved the following normalized relevance: 100% (uric acid), 10.3% (creatinine), 9.8% (AST), 5.4% (lymphocytes), and 5% (C-reactive protein) for having joint pain. Uric acid has the highest normalized relevance for predicting joint pain. The earliest artificial intelligence-based detection of joint pain will aid in the prevention of more serious orthopedic complications.