Random Forest Classification Model Research Articles

Selection of the level of vibration signal decomposition and mother wavelets to determine the level of failure severity in spur gearboxes

AbstractSpur gearboxes are an integral component in the operation of rotary machines. Hence, the early determination of the severity level of a failure is crucial. This manuscript delineates a methodology for selecting essential mother wavelets and filters from the wavelet transform (WT) to process the vibration signal within the time‐frequency domain, aiming to ascertain the severity level of failures in spur gearboxes. Initially, information is garnered from the gearbox through vibration signals in the time domain, utilising six accelerometers. Subsequently, the signal is partitioned into various levels, and information from each level is extracted using diverse mother wavelets and their respective filters. The signal is segmented into sub‐bands, from which the condition state is ascertained using an energy operator. After that, the appropriate level of wave decomposition is determined through ANOVA tests and post‐hoc Tukey analyses, evaluating performance in failure classification via the Random Forest (RF) model. Upon establishing the decomposition level, the analysis proceeds to identify which mother wavelets and filters are most suitable for determining the severity level of different types of failure in spur gearboxes. Moreover, this study investigates the impact of sensor positioning and inclination on acquiring the vibration signal. This aspect is explored through factorial ANOVA tests and multiple comparisons of the data derived from the sensors. The RF classification model achieved exceedingly favourable results (accuracy 96% and AUC 98%), with minimal practical influence from the positioning and inclination of a sensor, thereby affirming the proposed methodology's suitability for this type of analysis.

Random forest predictive modeling of prolonged hospital length of stay in elderly hip fracture patients

BackgroundIn elderly individuals suffering from hip fractures, a prolonged hospital length of stay (PLOS) not only heightens the probability of patient complications but also amplifies mortality risks. Yet, most elderly hip fracture patients present compromised baseline health conditions. Additionally, PLOS leads to increased expenses for patient treatment and care, while also diminishing hospital turnover rates. This, in turn, jeopardizes the prompt allocation of beds for urgent cases.MethodsA retrospective study was carried out from October 2021 to November 2023 on 360 elderly hip fracture patients who underwent surgical treatment at West China Hospital. The 75th percentile of the total patient cohort’s hospital stay duration, which was 12 days, was used to define prolonged hospital length of stay (PLOS). The cohort was divided into training and testing datasets with a 70:30 split. A predictive model was developed using the random forest algorithm, and its performance was validated and compared with the Lasso regression model.ResultsOut of 360 patients, 103 (28.61%) experienced PLOS. A Random Forest classification model was developed using the training dataset, identifying 10 essential variables. The Random Forest model achieved perfect performance in the training set, with an area under the curve (AUC), balanced accuracy, Kappa value, and F1 score of 1.000. In the testing set, the model’s performance was assessed with an AUC of 0.846, balanced accuracy of 0.7294, Kappa value of 0.4325, and F1 score of 0.6061.ConclusionThis study aims to develop a prognostic model for predicting delayed discharge in elderly patients with hip fractures, thereby improving the accuracy of predicting PLOS in this population. By utilizing machine learning models, clinicians can optimize the allocation of medical resources and devise effective rehabilitation strategies for geriatric hip fracture patients. Additionally, this method can potentially improve hospital bed turnover rates, providing latent benefits for the healthcare system.

The distribution of global tidal marshes from Earth observation data

AbstractAimTidal marsh ecosystems are heavily impacted by human activities, highlighting a pressing need to address gaps in our knowledge of their distribution. To better understand the global distribution and changes in tidal marsh extent, and identify opportunities for their conservation and restoration, it is critical to develop a spatial knowledge base of their global occurrence. Here, we develop a globally consistent tidal marsh distribution map for the year 2020 at 10‐m resolution.LocationGlobal.Time period2020.Major taxa studiedTidal marshes.MethodsTo map the location of the world's tidal marshes at 10‐m resolution, we applied a random forest classification model to Earth observation data from the year 2020. We trained the classification model with a reference dataset developed to support distribution mapping of coastal ecosystems, and predicted the spatial distribution of tidal marshes between 60° N and 60° S. We validated the tidal marsh map using standard accuracy assessment methods, with our final map having an overall accuracy score of 0.85.ResultsWe estimate the global extent of tidal marshes in 2020 to be 52,880 km2 (95% CI: 32,030 to 59,780 km2) distributed across 120 countries and territories. Tidal marsh distribution is centred in temperate and Arctic regions, with nearly half of the global extent of tidal marshes occurring in the temperate Northern Atlantic (45%) region. At the national scale, over a third of the global extent (18,510 km2; CI: 11,200–20,900) occurs within the USA.Main conclusionsOur analysis provides the most detailed spatial data on global tidal marsh distribution to date and shows that tidal marshes occur in more countries and across a greater proportion of the world's coastline than previous mapping studies. Our map fills a major knowledge gap regarding the distribution of the world's coastal ecosystems and provides the baseline needed for measuring changes in tidal marsh extent and estimating their value in terms of ecosystem services.

A Qualitative Study of Water Quality Using Landsat 8 and Station Water Quality-Monitoring Data to Support SDG 6.3.2 Evaluations: A Case Study of Deqing, China

Facing the challenge of the degradation of global water quality, it is urgent to realize the Sustainable Development Goal 6.3.2 (SDG 6.3.2), which focuses on improving global water quality. Currently, remote sensing technology is widely used for water quality monitoring. Existing water quality-monitoring studies have been conducted based on quantitative water quality inversion. It requires a high degree of the synchronization of the time and location of the collection of station monitoring data and remote sensing data (air–ground spatiotemporal synchronization), which can be resource intensive and time consuming. However, policymakers and the public are more interested in the quality of water (good or poor) than in the specific values of the water quality parameters, as evidenced by the emergence of SDG 6.3.2. In this study, we change the traditional idea of quantitative water quality research, focus on water quality qualitative research combined with the characteristics of water pollution, propose a remote sensing water quality sample enhancement method under the condition of “air–ground spatiotemporal asynchrony”, and construct a remote sensing water quality sample library. On the basis of this sample library, a random forest water quality classification model was constructed to classify water quality qualitatively. We obtained the distribution of good water bodies in Deqing County, China, for example, from 2013 to 2022. The results show that the model has high accuracy (Kappa = 0.6004, OA = 0.8387), and we found that the water quality in Deqing County improved in the order of “major rivers, lakes, and tributaries” during the period from 2013 to 2015. This also verifies the feasibility of using this sample enhancement method to conduct qualitative research on water quality. Based on this water quality classification model, a set of spatial-type evaluation processes of SDG 6.3.2 based on image elements was designed. The evaluation results show that the water quality situation in Deqing County can be divided into two stages: there is a trend of substantial improvement from 2013 (evaluated value of SDG 6.3.2 = 63.25) to 2015 (evaluated value of SDG 6.3.2 = 83.16); and it has remained stable and fluctuating after reaching the good environmental water quality since 2015. This study proposes a simple method for rapidly evaluating SDG 6.3.2 via utilizing easily accessible Landsat 8 and water quality-monitoring data to classify water quality. The method can directly obtain water quality category information without the need for additional sampling, thus saving costs. It is a very simple process that is easy to implement, while also providing a high level of accuracy. This significantly reduces the barriers to evaluating SDG 6.3.2, supports the realization of the sustainable management of water resources globally, and is highly generalizable.

Improved random forest classification model combined with C5.0 algorithm for vegetation feature analysis in non-agricultural environments

In response to the challenges posed by the high computational complexity and suboptimal classification performance of traditional random forest algorithms when dealing with high-dimensional and noisy non-agricultural vegetation satellite data, this paper proposes an enhanced random forest algorithm based on the C5.0 algorithm. The paper focuses on the Liaohe Plain, selecting two distinct non-agricultural landscape patterns in Shenbei New District and Changtu County as research objects. High-resolution satellite data from GF-2 serves as the experimental dataset. This paper introduces an ensemble feature method based on the bagging concept to improve the original random forest classification model. This method enhances the likelihood of selecting features beneficial to classifying positive class samples, avoiding excessive removal of useful features from negative samples. This approach ensures feature importance and model diversity. The C5.0 algorithm is then employed for feature selection, and the enhanced vegetation index (EVI) is utilized for vegetation coverage estimation. Results indicate that employing a multi-scale parameter selection tool, combined with limited field-measured data, facilitates the identification and classification of plant species in forest landscapes. The C5.0 algorithm effectively selects classification features, minimizing information redundancy. The established object-oriented random forest classification model achieves an impressive accuracy of 94.02% on the aerial imagery for forest classification dataset, with EVI-based vegetation coverage estimation demonstrating high accuracy. In experiments on the same test set, the proposed algorithm attains an average accuracy of 90.20%, outperforming common model algorithms such as bidirectional encoder representation from transformer, FastText, and convolutional neural network, which achieve average accuracies ranging from 84.41 to 88.33% in identifying non-agricultural artificial habitat vegetation features. The proposed algorithm exhibits a competitive edge compared to other algorithms. These research findings contribute scientific evidence for protecting agricultural ecosystems and restoring agricultural ecosystem biodiversity.

Abstract PO3-07-04: Prediction of PAM50 molecular subtypes from H&E-stained breast cancer specimens using tumor microenvironment features and additive multiple instance learning models

Abstract Background: PAM50, a 50-gene signature, classifies breast cancers into one of five subtypes (basal, luminal A, luminal B, HER2-enriched, and normal-like), revealing information about underlying tumor biology, and has emerged as a key prognostic indicator influencing treatment decisions. There is growing interest in bridging the gap between expression-based metrics and histopathology, where immunohistochemistry (IHC) and sequencing-based approaches have been proposed for this purpose. However, hematoxylin and eosin (H&E)-stained slides are ubiquitously utilized by pathologists for cancer diagnosis, while IHC and sequencing-based approaches require additional tissue and specialized processing and/or analysis. Here, we describe a computer vision-based approach to predict PAM50 classification using H&E-stained whole slide images (WSIs). Methods: We obtained expression-based PAM50 subtype labels and corresponding H&E-stained WSIs for 961 breast carcinomas from the TCGA BRCA cohort. We used two separate machine learning (ML) approaches to predict PAM50 subtypes from WSIs. In the first approach, we deployed previously trained PathExplore models to extract quantitative human-interpretable features (HIFs) that summarize the TME. We subsequently trained random forest classification models on these HIFs to predict PAM50 subtypes. For the second approach, we developed additive multiple instance learning (aMIL) models. Additionally, we explored the effects of PAM50 subtype labeling and aggregation strategies beyond the 5-class approach. Our 3-class approach combines Luminal A and B, as seen in IHC efforts to increase agreement with PAM50 assays, while excluding Normal, a category containing few and heterogeneous samples. We also performed binary classification for each subtype in the 3-class model (e.g. luminal vs. other). Slides were split into training (60%), validation (20%), and test (20%) sets, stratified by PAM50 labels, and model performance was assessed using the area under the receiver operator curve (AUROC) metric on the held-out test set, using a one vs. rest approach for multi-class models. To establish a baseline for PAM50 prediction, we developed random forest classification models using only clinical covariates (tumor stage, histologic grade, histological subtype, and BRCA1/2 status). Results: We compared the performance of our two ML models (HIF and aMIL) to that of the baseline model, and we report the AUROC values in Table 1. These models both performed well in predicting Basal, Luminal A, Luminal B, and Luminal (A+B), while the model performance was less strong for predictions of the HER2 and Normal classifications. The three-class model showed improved performance of predicting Luminal classifications relative to the five-class model that separates Luminal A and B. Although simplifying classification problems to a binary use case typically provides improved performance, this phenomenon was not observed for any of the PAM50 subtypes. Conclusions: These results demonstrate that AI-powered digital pathology can accurately and reproducibly perform molecular-based classification tasks, such as predicting PAM50 classifications, using WSIs, suggesting a more efficient path toward clinically relevant breast cancer characterization. Table 1. Performance of all models in predicting PAM50 molecular subtypes. AUROC values are shown. Shaded cells represent the best test-set performance for each class (row). Citation Format: Maria Guramare, Syed Ashar Javed, Christian Kirkup, Dinkar Juyal, Jacqueline Brosnan-Cashman, Victoria Mountain, Ryan Leung, Bahar Rahsepar, John Abel, Amaro Taylor-Weiner, Jake Conway. Prediction of PAM50 molecular subtypes from H&E-stained breast cancer specimens using tumor microenvironment features and additive multiple instance learning models [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO3-07-04.

Three-dimensional mineral prospectivity mapping based on natural language processing and random forests: A case study of the Xiyu diamond deposit, China

The advent of the big data era has gradually brought about a new demand for the integration of geological text big data in three-dimensional (3D) mineral prospectivity mapping (3DMPM). Here, we report a novel workflow suitable for the application of natural language processing in 3DMPM. Our study area is the Xiyu kimberlite-type diamond deposit in Shandong, China, which has significant prospecting potential in its deep portion. In this study, exploration criteria served as a bridge connecting geological text big data and 3DMPM, and exploration criteria for the Xiyu diamond deposit were constructed employing text mining technologies. A comparative evaluation of conventional and text-mining-based exploration criteria was also conducted. Considering the text-mining-based deposit exploration criteria, we reconstructed 3D spatial anomalies and quantitatively analyzed ore-controlling factors. A random forests classification model, created using the deposit exploration criteria obtained through text mining, was applied to prospecting predictions in the Xiyu deposit, with superior results being obtained. The prediction results confirmed the applicability of the workflow and demonstrated its capacity for effective dual coupling of geological data and knowledge.

Inflammatory Protein Signatures as Predictive Disease-Specific Markers for Non-Alcoholic Steatohepatitis (NASH).

Non-alcoholic fatty liver disease (NAFLD) is the most prevalent chronic disease worldwide, consisting of a broad spectrum of diseases such as simple steatosis (NAFL), non-alcoholic steatohepatitis (NASH), fibrosis, cirrhosis, and hepatocellular carcinoma. Hepatic inflammation plays a key role in the pathophysiology of NAFLD. Inflammatory mediators such as cytokines and chemokines are considered as contributing factors to NAFLD development and progression. In the present study, we aimed to investigate the inflammatory protein signatures as predictive disease-specific markers for non-alcoholic fatty liver disease (NAFLD). This cross-sectional study included healthy control (n = 64), NAFL (n = 109), and NASH (n = 60) human subjects. Serum concentrations of various cytokines and chemokines were evaluated using sensitive multiplex assays. We used principal component analysis (PCoA) to reveal distinct differences in the levels of cytokines and chemokines between each of the study groups. Further, a random forest classification model was developed to identify the panel of markers that could predict diseases. The protein-protein network analysis was performed to determine the various signaling pathways associated with the disease-specific panel of markers. Serum concentrations of TNF-α, IL-1β, IL-1ra, G-CSF, PDGF-BB, MCP-1, MIP-1a, MIP-1b, RANTES, eotaxin, IL-8 and IP-10 were significantly increased in NASH group as compared to control group. Furthermore, serum concentrations of IL-9 and IL-13 were significantly lower in the NASH group, whereas IL-2 levels were significantly decreased in the NAFL group when compared to the control group. PCoA results demonstrated statistically significant differences in cytokines and chemokines between each of the study groups (PERMANOVA p = 0.001; R2 = 0.102). RANTES, IL-1ra, MIP-1b, IL-2, and G-CSF could differentiate the NAFL group from the controls; G-CSF, IL-1ra, TNF-α, RANTES, and IL-9 could differentiate the NASH group from the controls; and G-CSF, IL-9, IL-13, eotaxin, and TNF- α could differentiate the NASH group from the NAFL group. Our protein-protein network revealed that these markers are involved in cytokine-cytokine receptor interaction, Th1 and Th2 cell differentiation, TNF, chemokine, JAK/STAT, P13K/Akt, TLR, NOD-like receptor, NF-kB, and adipocytokine signaling pathways which might be responsible for disease pathogenesis.Our study findings revealed a set of distinct cytokine and chemokine markers and they might be considered as biomarkers in distinguishing NASH from NAFL. Future multicentre studies with larger sample size are recommended to determine the potential utility of these panels of markers.

Detection of pine wood nematode infestation using hyperspectral drone images

The pine wood nematode (PWN) is one of the most threatening pests of pine forests in China. An essential management response involves felling and removing the infested trees before the PWN can spread over large areas, but this requires early detection. In this study, we analyzed the temporal changes in the spectral signatures of healthy and infested trees throughout the entire growing season (May to October) using unmanned aerial vehicle (UAV) hyperspectral images (HSI). To explore the potential of these seasonal spectral changes in detecting PWN, we investigated the separability of healthy and infested trees using both absolute values and changes in values. A random forest (RF) classifier was applied to separate healthy and infested trees based on reflectances, derivatives, vegetation indices (VIs), and their seasonal change values. We obtained the following results: (1) the seasonal changes of tree crown spectra following a PWN infestation were similar to those observed using absolute spectral values, though the latter performed better in early detection; (2) in the early stage of infestation (June and July), the green and red-edge regions contained the most sensitive bands, whereas the near-infrared portion played a more important role in later stages; (3) derivatives achieved higher overall accuracy (OA) than the reflectance in the early stage of PWN infestation (June and July); (4) in June, the Pigment Specific Simple Ratio (PSSR) had the highest separability (OA = 0.77) for PWN infestation, followed by the Plant Senescence Reflectance Index (PSRI) and the Normalized Difference Red-edge Index (NDRE), both with a OA of 0.70. In July, PSRI, Red Edge Normalized Difference Vegetation Index (RENDVI), and the Modified Red Edge Simple Ratio (MRESR) had the highest OA of 0.77. These VIs were calculated based on the red-edge regions. In August, RENDVI achieved the best accuracy (0.87), followed by the Normalized Difference Vegetation Index (NDVI, OA = 0.83) and the Plant Stress Index (PSI, OA = 0.80). In September and October, PSRI, PSI, the Greenness Index (GI), and RENDVI showed the strongest separability (0.87 to 0.90). Similar results were obtained using change values of VIs. In June, RENDVI, PSI, and NDVI showed the highest accuracy (OA = 0.73). In July, MRESR had the best accuracy (0.77), followed by RENDVI (0.73) and PSI (0.73). In the late stage of PWN infestation, NDVI, PSRI, PSI, RENDVI, GI, and PSSR obtained the highest accuracies (0.83 to 0.93), surpassing the accuracy achieved using absolute spectra; (5) in RF classification model, seasonal variation values derived from reflectances, derivatives, and VIs performed well in the late stages of the PWN infestation (September and October) but were not as good as absolute values in the early and middle stages (June to August). Our study explored the potential of seasonal changes in UAV spectra for early detection of forest pests.

Assessment of Risk Factors for Acute Kidney Injury with Machine Learning Tools in Children Undergoing Hematopoietic Stem Cell Transplantation.

Background: Although acute kidney injury (AKI) is a common complication in patients undergoing hematopoietic stem cell transplantation (HSCT), its prophylaxis remains a clinical challenge. Attempts at prevention or early diagnosis focus on various methods for the identification of factors influencing the incidence of AKI. Our aim was to test the artificial intelligence (AI) potential in the construction of a model defining parameters predicting AKI development. Methods: The analysis covered the clinical data of children followed up for 6 months after HSCT. Kidney function was assessed before conditioning therapy, 24 h after HSCT, 1, 2, 3, 4, and 8 weeks after transplantation, and, finally, 3 and 6 months post-transplant. The type of donor, conditioning protocol, and complications were incorporated into the model. Results: A random forest classifier (RFC) labeled the 93 patients according to presence or absence of AKI. The RFC model revealed that the values of the estimated glomerular filtration rate (eGFR) before and just after HSCT, as well as methotrexate use, acute graft versus host disease (GvHD), and viral infection occurrence, were the major determinants of AKI incidence within the 6-month post-transplant observation period. Conclusions: Artificial intelligence seems a promising tool in predicting the potential risk of developing AKI, even before HSCT or just after the procedure.

Evaluating a Novel Approach to Detect the Vertical Structure of Insect Damage in Trees Using Multispectral and Three-Dimensional Data from Drone Imagery in the Northern Rocky Mountains, USA

Remote sensing is a well-established tool for detecting forest disturbances. The increased availability of uncrewed aerial systems (drones) and advances in computer algorithms have prompted numerous studies of forest insects using drones. To date, most studies have used height information from three-dimensional (3D) point clouds to segment individual trees and two-dimensional multispectral images to identify tree damage. Here, we describe a novel approach to classifying the multispectral reflectances assigned to the 3D point cloud into damaged and healthy classes, retaining the height information for the assessment of the vertical distribution of damage within a tree. Drone images were acquired in a 27-ha study area in the Northern Rocky Mountains that experienced recent damage from insects and then processed to produce a point cloud. Using the multispectral data assigned to the points on the point cloud (based on depth maps from individual multispectral images), a random forest (RF) classification model was developed, which had an overall accuracy (OA) of 98.6%, and when applied across the study area, it classified 77.0% of the points with probabilities greater than 75.0%. Based on the classified points and segmented trees, we developed and evaluated algorithms to separate healthy from damaged trees. For damaged trees, we identified the damage severity of each tree based on the percentages of red and gray points and identified top-kill based on the length of continuous damage from the treetop. Healthy and damaged trees were separated with a high accuracy (OA: 93.5%). The remaining damaged trees were separated into different damage severities with moderate accuracy (OA: 70.1%), consistent with the accuracies reported in similar studies. A subsequent algorithm identified top-kill on damaged trees with a high accuracy (OA: 91.8%). The damage severity algorithm classified most trees in the study area as healthy (78.3%), and most of the damaged trees in the study area exhibited some amount of top-kill (78.9%). Aggregating tree-level damage metrics to 30 m grid cells revealed several hot spots of damage and severe top-kill across the study area, illustrating the potential of this methodology to integrate with data products from space-based remote sensing platforms such as Landsat. Our results demonstrate the utility of drone-collected data for monitoring the vertical structure of tree damage from forest insects and diseases.

Adaptive Random Forest for Gait Prediction in Lower Limb Exoskeleton

To improve the human-machine cooperativity of a wearable lower limb exoskeleton, a gait recognition method based on surface electromyography (sEMG) was proposed. sEMG of rectus femoris, vastus medialis, vastus lateralis, semitendinosus and biceps femoris were acquired. Then, time domain, frequency domain, time-frequency domain and nonlinear features were extracted. The integrated value of electromyography, variance, root mean square and wavelength were selected as the time domain features and the frequency domain feature includes mean power frequency. Wavelet packet energy was selected as the time-frequency domain feature. Nonlinear features including approximate entropy, sample entropy and fuzzy entropy of sEMG were extracted. Classification accuracy of different feature matrices and different muscle groups were constructed and verified. The optimal multi-dimensional fusion feature matrix was determined. Introducing the Bayesian optimization algorithm, the Bayesian optimized Random Forest classification model was constructed to identify different gait phases. Comparing with Random Forest, the accuracy of the optimized Random Forest was improved by 5.89%. Applying Random Forest algorithm with Bayesian optimization to gait prediction based on sEMG, the followership and consistency of gait control in lower limb exoskeleton can be improved. This template explains and demonstrates how to prepare your camera-ready paper for Trans Tech Publications. The best is to read these instructions and follow the outline of this text.

A RANDOM FOREST-BASED MODEL OF SCORE FLUCTUATIONS IN PROFESSIONAL TENNIS MATCHES

In tennis matches, the victory and turning points of the game are often influenced by various factors. To explore the factors that affect match fluctuations (changes in the flow of scoring) and to provide suggestions for athletes match strategies, this paper first identifies general indicators through a literature review and uses logistic regression to determine the effectiveness of the chosen model. Secondly, it employs the Fourier function fitting to identify turning points in the match. Considering the scarcity of turning points in the game, this paper uses the SMOTE method to expand the dataset. Subsequently, it tests with a random forest classification model, achieving an accuracy of 93.433%. To improve the models accuracy, several indicators were added to the original model, resulting in a correct rate of 98.51%. Finally, to verify the models results and applicability, the model was applied to other matches with good results. A sensitivity analysis was conducted, revealing good model stability. The model results indicate that the main factors affecting the appearance of turning points include the players movement distance during the match, whether there are changes in the depth and width of the return, score differences, and the maximum number of consecutive wins. When tested in other types of matches, we found that the importance of these factors may change to some extent, but the results remain satisfactory. KEYWORDS: Volatility prediction, Random Forest, Logistic regression, Sensitivity analysis.

Passive acoustic data yields insights into bird vocalization behavior associated with invasive shrub removal

AbstractMonitoring biodiversity changes associated with ecological restoration is crucial in the current UN Decade on Restoration. Although several studies highlight the impacts of restoration on ecosystems, it is also important to understand how restoration alters species' behavior, including vocalization. Advances in conservation technology, such as passive acoustic monitoring, facilitate rapid and noninvasive monitoring. In this study, we quantified changes in the vocal behavior of a bird species in response to small‐scale restoration carried out by removing the invasive shrub, Lantana camara (lantana), in a Central Indian tropical dry forest. We examined associations between sites of varying lantana densities and the vocalization of one of its primary dispersers, red‐vented bulbul (RVBU). We found statistically significant differences in note‐length and bandwidth of RVBU vocalizations across sites. A random forest classification model showed that Lantana density was not an important predictor of RVBU vocalizations. Apart from the percentage of forest area and farms in our sites, total human population was the most important predictor for RVBU vocalizations suggesting RVBU's use of human habitations as possible resource hubs. Our findings suggest that lantana removal is not associated with significant changes in RVBU vocalization. This study provides empirical evidence that ecological restoration may not alter species' vocalization in the short term and highlights the importance of moving beyond species presence and understanding the impact of restoration on species behavior.

Precipitation Estimation Using FY-4B/AGRI Satellite Data Based on Random Forest

Precipitation is the basic component of the Earth’s water cycle. Obtaining high-resolution and high-precision precipitation data is of great significance. This paper establishes a precipitation retrieval model based on a random forest classification and regression model during the day and at night with FY-4B/AGRI Level1 data on China from July to August 2022. To evaluate the retrieval effect of the model, the GPM IMERG product is used as a reference, and the retrieval results are compared against those of the FY-4B/AGRI operational precipitation product. In addition, the retrieval results are analyzed according to different underlying surfaces. The results showed that compared with the FY-4B/AGRI operational precipitation product, the retrieval model can better identify precipitation and capture precipitation areas of light rain, moderate rain, heavy rain and torrential rain. Among them, the probability of detection (POD) of the day model increased from 0.328 to 0.680, and the equitable threat score (ETS) increased from 0.252 to 0.432. The POD of the night model increased from 0.337 to 0.639, and the ETS score increased from 0.239 to 0.369. Meanwhile, the precipitation estimation accuracy of the day model increased by 38.98% and that of the night model increased by 40.85%. Our results also showed that due to the surface uniformity of the ocean, the model can identify precipitation better on the ocean than on the land. Our findings also indicated that for the different underlying surfaces of the land, there is no significant difference in each evaluation index of the model. This is a strong argument for the universal applicability of the model. Notably, the results showed that, especially for more vegetated areas and areas covered by water, the model is capable of estimating precipitation. In conclusion, the precipitation retrieval model that is proposed herein can better determine precipitation regions and estimate precipitation intensities compared with the FY-4B/AGRI operational precipitation product. It can provide some reference value for future precipitation retrieval research on FY-4B/AGRI.

Accurate species classification of Arctic toothed whale echolocation clicks using one-third octave ratios.

Passive acoustic monitoring has been an effective tool to study cetaceans in remote regions of the Arctic. Here, we advance methods to acoustically identify the only two Arctic toothed whales, the beluga (Delphinapterus leucas) and narwhal (Monodon monoceros), using echolocation clicks. Long-term acoustic recordings collected from moorings in Northwest Greenland were analyzed. Beluga and narwhal echolocation signals were distinguishable using spectrograms where beluga clicks had most energy >30 kHz and narwhal clicks had a sharp lower frequency limit near 20 kHz. Changes in one-third octave levels (TOL) between two pairs of one-third octave bands were compared from over one million click spectra. Narwhal clicks had a steep increase between the 16 and 25 kHz TOL bands that was absent in beluga click spectra. Conversely, beluga clicks had a steep increase between the 25 and 40 kHz TOL bands that was absent in narwhal click spectra. Random Forest classification models built using the 16 to 25 kHz and 25 to 40 kHz TOL ratios accurately predicted the species identity of 100% of acoustic events. Our findings support the use of echolocation TOL ratios in future automated click classifiers for acoustic monitoring of Arctic toothed whales and potentially for other odontocete species.

Enhancing prediction and analysis of UK road traffic accident severity using AI: Integration of machine learning, econometric techniques, and time series forecasting in public health research

This research project explored into the intricacies of road traffic accidents severity in the UK, employing a potent combination of machine learning algorithms, econometric techniques, and traditional statistical methods to analyse longitudinal historical data. Our robust analysis framework includes descriptive, inferential, bivariate, multivariate methodologies, correlation analysis: Pearson's and Spearman's Rank Correlation Coefficient, multiple logistic regression models, Multicollinearity Assessment, and Model Validation. In addressing heteroscedasticity or autocorrelation in error terms, we've advanced the precision and reliability of our regression analyses using the Generalized Method of Moments (GMM). Additionally, our application of the Vector Autoregressive (VAR) model and the Autoregressive Integrated Moving Average (ARIMA) models have enabled accurate time series forecasting. With this approach, we've achieved superior predictive accuracy and marked by a Mean Absolute Scaled Error (MASE) of 0.800 and a Mean Error (ME) of -73.80 compared to a naive forecast. The project further extends its machine learning application by creating a random forest classifier model with a precision of 73%, a recall of 78%, and an F1-score of 73%. Building on this, we employed the H2O AutoML process to optimize our model selection, resulting in an XGBoost model that exhibits exceptional predictive power as evidenced by an RMSE of 0.1761205782994506 and MAE of 0.0874235576229789. Factor Analysis was leveraged to identify underlying variables or factors that explain the pattern of correlations within a set of observed variables. Scoring history, a tool to observe the model's performance throughout the training process was incorporated to ensure the highest possible performance of our machine learning models. We also incorporated Explainable AI (XAI) techniques, utilizing the SHAP (Shapley Additive Explanations) model to comprehend the contributing factors to accident severity. Features such as Driver_Home_Area_Type, Longitude, Driver_IMD_Decile, Road_Type, Casualty_Home_Area_Type, and Casualty_IMD_Decile were identified as significant influencers. Our research contributes to the nuanced understanding of traffic accident severity and demonstrates the potential of advanced statistical, econometric, machine learning techniques in informing evidence based interventions and policies for enhancing road safety.

Utilizing Machine Learning to Forecast Smoking Behavior

Abstract: The examination of smoking habits, extensively studied over time, has posed challenges in accurately predicting and thoroughly analyzing its determinants. Previous research efforts struggled to precisely predict smoking behavior due to the presence of continuous target variables, which hindered the application of crucial feature selection techniques like mutual information. This study aims to tackle these hurdles through an innovative approach that integrates data preprocessing, feature engineering, and advanced machine learning methods. To address the issue of continuous target variables, our methodology involves categorizing smoking behavior into discrete groups, enabling the utilization of feature selection techniques such as mutual information scores. Logistic regression, Gaussian Naive Bayes, and Random Forest Classifier models are utilized to achieve highly accurate predictions of smoking behavior. The Select KBest method is employed to evaluate the importance of features based on mutual information scores. The investigation delves into various health markers, including BMI, haemoglobin levels, and cholesterol, offering comprehensive insights into their influence on smoking habits. Furthermore, Principal Component Analysis (PCA) is implemented to reduce dimensionality while preserving essential information from the dataset. Through this novel approach and a steadfast commitment to ethical data collection practices, our objective is to advance the comprehension of smoking behavior, surmounting past challenges, and providing valuable insights for public health initiatives and smoking cessation endeavors. The paper assesses outcomes using specified algorithms and parameters, presenting a comparative analysis to enhance the clarity and reliability of our findings.

Study of prognostic splicing factors in cancer using machine learning approaches.

Splicing factors (SFs) are the major RNA-binding proteins (RBPs) and key molecules that regulate the splicing of mRNA molecules through binding to mRNAs. The expression of splicing factors is frequently deregulated in different cancer types, causing the generation of oncogenic proteins involved in cancer hallmarks. In this study, we investigated the genes that encode RNA-binding proteins and identified potential splicing factors that contribute to the aberrant splicing applying a random forest classification model. The result suggested 56 splicing factors were related to the prognosis of 13 cancers, two SF complexes in liver hepatocellular carcinoma, and one SF complex in esophageal carcinoma. Further systematic bioinformatics studies on these cancer prognostic splicing factors and their related alternative splicing events revealed the potential regulations in a cancer-specific manner. Our analysis found high ILF2-ILF3 expression correlates with poor prognosis in LIHC through alternative splicing. These findings emphasize the importance of SFs as potential indicators for prognosis or targets for therapeutic interventions. Their roles in cancer exhibit complexity and are contingent upon the specific context in which they operate. This recognition further underscores the need for a comprehensive understanding and exploration of the role of SFs in different types of cancer, paving the way for their potential utilization in prognostic assessments and the development of targeted therapies.

Data engineering for predictive machine learning of stormwater infrastructure conditions

Culverts are stormwater structures that require ongoing condition assessment and proactive maintenance for optimal performance to reduce flooding, especially in urban areas. Managing these structures can be complex and costly regarding time, money, and personnel resources. Machine learning (ML) tools are a powerful means of predicting culvert conditions via learning from their existing data records. However, the data available on such subjects are usually low in quantity and of various qualities; thus, they need significant data processing for ML. Existing stormwater infrastructure condition prediction studies rarely detail their data processing methods. Hence, this work uses a comprehensive case study to illustrate essential data preparation techniques to address common data issues and enhance the performance of commonly used ML algorithms. The study shows methods for exploratory data analysis to understand the dataset, data wrangling methods for preprocessing data of various quality, and data engineering procedures for addressing insufficient data issues. After the data processing procedures are applied, F-1 scores are used to evaluate the ML models' performance. The random forest classifier model, one of the four ML models, performed best after applying the data wrangling and data engineering methods. Overall, this work provides transparency of methods and applications to encourage ML use in the water resources engineering field.