Random Forest Algorithm Research Articles

Estimating Grassland Carrying Capacity in the Source Area of Nujiang River and Selinco Lake, Tibetan Plateau (2001–2020) Based on Multisource Remote Sensing

Estimating the spatiotemporal variations in natural grassland carrying capacity is crucial for maintaining the balance between grasslands and livestock. However, accurately assessing this capacity presents significant challenges due to the high costs of biomass measurement and the impact of human activities. In this study, we propose a novel method to estimate grassland carrying capacity based on potential net primary productivity (NPP), applied to the source area of the Nujiang River and Selinco Lake on the Tibetan Plateau. Initially, we utilize multisource remote sensing data—including soil, topography, and climate information—and employ the random forest regression algorithm to model potential NPP in areas where grazing is banned. The construction of the random forest model involves rigorous feature selection and hyperparameter optimization, enhancing the model’s accuracy. Next, we apply this trained model to areas with grazing, ensuring a more accurate estimation of grassland carrying capacity. Finally, we analyze the spatiotemporal variations in grassland carrying capacity. The main results showed that the model achieved a high level of precision, with a root mean square error (RMSE) of 4.89, indicating reliable predictions of grassland carrying capacity. From 2001 to 2020, the average carrying capacity was estimated at 9.44 SU/km2, demonstrating a spatial distribution that decreases from southeast to northwest. A slight overall increase in carrying capacity was observed, with 65.7% of the area exhibiting an increasing trend, suggesting that climate change has a modest positive effect on the recovery of grassland carrying capacity. Most of the grassland carrying capacity is found in areas below 5000 m in altitude, with alpine meadows and alpine meadow steppes below 4750 m being particularly suitable for grazing. Given that the overall grassland carrying capacity remains low, it is crucial to strictly control local grazing intensity to mitigate the adverse impacts of human activities. This study provides a solid scientific foundation for developing targeted grassland management and protection policies.

The PANoptosis-related hippocampal molecular subtypes and key biomarkers in Alzheimer’s disease patients

Alzheimer’s Disease (AD) is a neurodegenerative disorder, and various molecules associated with PANoptosis are involved in neuroinflammation and neurodegenerative diseases. This work aims to identify key genes, and characterize PANoptosis-related molecular subtypes in AD. Moreover, we establish a scoring system for distinguishing PANoptosis molecular subtypes and constructing diagnostic models for AD differentiation. A total of 5 hippocampal datasets were obtained from the Gene Expression Omnibus (GEO) database. In total, 1324 protein-encoding genes associated with PANoptosis (1313 apoptosis genes, 11 necroptosis genes, and 31 pyroptosis genes) were extracted from the GeneCards database. The Limma package was used to identify differentially expressed genes. Weighted Gene Co-Expression Network Analysis (WGCNA) was conducted to identify gene modules significantly associated with AD. The ConsensusClusterPlus algorithm was used to identify AD subtypes. Gene Set Variation Analysis (GSVA) was used to assess functional and pathway differences among the subtypes. The Boruta, Least Absolute Shrinkage and Selection Operator (LASSO), Random Forest (RF), and Support Vector Machine Recursive Feature Elimination (SVM-RFE) algorithms were used to select the three PANoptosis-related Key AD genes (PKADg). A scoring model was constructed based on the Boruta algorithm. PANoptosis diagnostic models were developed using the RF, SVM-RFE, and Logistic Regression (LR) algorithms. The ROC curves were used to assess the model performance. A total of 48 important genes were identified by intersecting 725 differentially expressed genes and 2127 highly correlated module genes from WGCNA with 1324 protein-encoding genes related to PANoptosis. Machine learning algorithms identified 3 key AD genes related to PANoptosis, including ANGPT1, STEAP3, and TNFRSF11B. These genes had strong discriminatory capacities among samples, with Receiver Operating Characteristic Curve (ROC) analysis indicating Area Under the Curve (AUC) values of 0.839, 0.8, and 0.868, respectively. Using the 48 important genes, the ConsensusClusterPlus algorithm identified 2 PANoptosis subtypes among AD patients, i.e., apoptosis subtype and mild subtype. Apoptosis subtype patients displayed evident cellular apoptosis and severe functionality damage in the hippocampal tissue. Meanwhile, mild subtype patients showed milder functionality damage. These two subtypes had significant differences in apoptosis and necroptosis; however, there was no apparent variation in pyroptosis functionality. The scoring model achieved an AUC of 100% for sample differentiation. The RF PANoptosis diagnostic model demonstrated an AUC of 100% in the training set and 85.85% in the validation set for distinguishing AD. This study identified two PANoptosis-related hippocampal molecular subtypes of AD, identified key genes, and established machine learning models for subtype differentiation and discrimination of AD. We found that in the context of AD, PANoptosis may influence disease progression through the modulation of apoptosis and necrotic apoptosis.

Land use land cover change detection in Baghdad city for the years 2000 and 2024 using random forest algorithm

Land use land cover change detection in Baghdad city for the years 2000 and 2024 using random forest algorithm

Prediction of allograft function in pre-transplant kidneys using sound touch elastography (STE): an ex vivo study

BackgroundThe purpose of the study was to evaluate renal quality and predict posttransplant graft function using ex vivo sound touch elastography (STE).MethodsIn this prospective study, 106 donor kidneys underwent ex vivo STE examination and biopsy from March 2022 to August 2023. The mean stiffness of the superficial cortex (STEsc), deep cortex (STEdc), and medulla (STEme) was obtained and synthesized into one index (STE) through the factor analysis method. Additionally, 100 recipients were followed up for 6 months. A random forest algorithm was employed to explore significant predictive factors associated with the Remuzzi score and allograft function. The performance of parameters was evaluated by using the area under the receiver operating characteristic curve (AUC).ResultsSTE had AUC values of 0.803 for diagnosing low Remuzzi and 0.943 for diagnosing high Remuzzi. Meanwhile, STE had an AUC of 0.723 for diagnosing moderate to severe ATI. Random forest algorithm identified STE and Remuzzi score as significant predictors for 6-month renal function. The AUC for STE in predicting postoperative allograft function was 0.717, which was comparable with that of the Remuzzi score (AUC = 0.756). Nevertheless, the specificity of STE was significantly higher than that of Remuzzi (0.913 vs 0.652, p < 0.001). Given these promising results, donor kidneys can be transplanted directly without the need for biopsy when STE ≤ 11.741.ConclusionsThe assessment of kidney quality using ex vivo STE demonstrated significant predictive value for the Remuzzi score and allograft function, which could help avoid unnecessary biopsy.Critical relevance statementPre-transplant kidney quality measured with ex vivo STE can be used to assess donor kidney quality and avoid unnecessary biopsy.Key PointsSTE has significant value for diagnosing low Remuzzi and high Remuzzi scores.STE achieved good performance in predicting posttransplant allograft function.Assessment of kidney quality using ex vivo STE could avoid unnecessary biopsies.Graphical

Flash flood prediction modeling in the hilly regions of Southeastern Bangladesh: A machine learning attempt on present and future climate scenarios

Flash floods are highly destructive, and their frequency and intensity are expected to escalate due to climatic changes. This study thus investigated flash flood susceptibility (FFS) by applying machine learning algorithms and climate projection to predict both present and future hazard scenarios in the southeastern hilly regions of Bangladesh. To predict FFS, we evaluated twelve flood-influencing variables: elevation (EL), slope (SL), aspect (AS), drainage density (DD), distance to stream (DS), topography roughness index (TRI), stream power index (SPI), topographic wetness index (TWI), soil permeability (SP), precipitation (PR), land use and land cover (LULC) and normalized difference vegetation index (NDVI). Earth observation data, field surveys, and past flood records were used to create a detailed flood inventory. Among the machine learning models tested, the random forest (RF) algorithm outperformed others, including support vector machine (SVC), logistic regression (LR), and extreme gradient boosting (XGBoost), and was subsequently used for flood susceptibility mapping based on future precipitation projections under two Sixth Coupled model intercomparison project (CMIP6) climate change scenarios: SSP1-2.6 and SSP5-8.5. Our findings indicated that the areas at high to very high risk of flooding are projected to increase significantly under both the SSP1-2.6 and SSP5-8.5 scenarios. Initially, around 38 % of the studied region had high to very high flood susceptibility, but this is expected to rise to 40–42 % over the projected time periods. These spatial delineations of flood-prone areas can provide guidance for developing effective mitigation and adaptation strategies to address the adverse impacts of flash flooding in the hilly river basins of Bangladesh.

Machine learning-aided microRNA discovery for olive oil quality.

MicroRNAs (miRNAs) are key regulators of gene expression in plants, influencing various biological processes such as oil quality and seed development. Although, our knowledge about miRNAs in olive (Olea europaea L.) is progressing, with several miRNAs being identified in previous studies, but most of these reported miRNAs have been predicted without the aid of a reference genome, primarily due to limited genome accessibility at the time. However, significant knowledge gaps still need to be improved in this area. This study addresses the complexities of miRNA detection in olive, using a high quality reference genome and a combination of genomics and machine learning-based methods. By leveraging random forest and support vector machine algorithms, we successfully identified 56 novel miRNAs in olive, surpassing the limitations of conventional homology-based methods. Our subsequent analysis revealed that some of these miRNAs are implicated in the regulation of key genes involved in oil quality. Within the context of oil biosynthesis pathways, the novel miRNA Oeu124369 regulates fatty acid biosynthesis by targeting acetyl-CoA acyltransferase 1 and palmitoyl-protein thioesterase, thereby influencing the production of acetyl-CoA and palmitic acid, respectively. These findings underscore the power of machine learning in unraveling the complex miRNA regulatory network in olive and provide a high quality miRNA resource for future research aimed at improving olive oil production by exploring the target genes of the identified miRNAs to understand their role and their biological processes.

Prediction of Ionospheric Scintillations Using Machine Learning Techniques during Solar Cycle 24 across the Equatorial Anomaly

Ionospheric scintillation is a pressing issue in space weather studies due to its diverse effects on positioning, navigation, and timing (PNT) systems. Developing an accurate and timely prediction model for this event is crucial. In this work, we developed two machine learning models for the prediction of ionospheric scintillation events at the equatorial anomaly during the maximum and minimum phases of solar cycle 24. The models developed in this study are the Random Forest (RF) algorithm and the eXtreme Gradient Boosting (XGBoost) algorithm. The models take inputs based on the solar wind parameters obtained from the OMNI Web database from the years 2010–2017 and Pc5 wave power obtained from the Bear Island (BJN) magnetometer station. We retrieved data from the Scintillation Network and Decision Aid (SCINDA) receiver in Egypt from which the S4 index was computed to quantify amplitude scintillations that were utilized as the target in the model development. Out-of-sample model testing was performed to evaluate the prediction accuracy of the models on unseen data after training. The similarity between the observed and predicted scintillation events, quantified by the R2 score, was 0.66 and 0.74 for the RF and XGBoost models, respectively. The corresponding Root Mean Square Errors (RMSEs) associated with the models were 0.01 and 0.01 for the RF and XGBoost models, respectively. The similarity in error shows that the XGBoost model is a good and preferred choice for the prediction of ionospheric scintillation events at the equatorial anomaly. With these results, we recommend the use of ensemble learning techniques for the study of the ionospheric scintillation phenomenon.

Machine Learning and Cointegration for Wind Turbine Monitoring and Fault Detection: From a Comparative Study to a Combined Approach

Data-driven models have become powerful tools for structural and condition monitoring of engineering systems, particularly wind turbines. This paper presents a comparative analysis of common machine learning (ML) algorithms (artificial neural networks, linear regression, random forests, and gradient boosting) and a cointegration-based approach for fault detection using Supervisory Control and Data Acquisition (SCADA) data. While ML models offer early fault prediction, the cointegration method is simpler, requires less training data, and has lower computational costs. However, it is less effective for early detection. To balance these trade-offs, we propose a cascading monitoring framework, where the ML model provides long-term predictions (outer monitoring process) and the cointegration model offers short-term verification (inner monitoring process). The cointegration model serves to confirm anomalies flagged by the ML model. By combining both models in a cascade structure, the system reduces the risk of false alarms triggered by uncertainties in the ML model alone. Furthermore, the short-term cointegration-based prediction model helps pinpoint immediate risks and mitigate the issue of prolonged downtime. This combination enhances both accuracy and reliability, as demonstrated through testing on a five-year SCADA dataset from a commercial wind turbine with a known gearbox fault.

Isolation Forest Anomaly Detection Algorithm Based On Multi-level Sub-subspace Partition

In the research of credit loan fraud detection, the isolation forest algorithm has attracted much attention because of its ability to efficiently process large-scale data sets. However, when facing high-dimensional data, the performance of the isolation forest algorithm is easily affected, resulting in deviation of the detection results. In order to solve the above problems, this paper proposes an isolation forest anomaly detection algorithm based on multi-level sub-subspace division. Firstly, the random forest algorithm is used to evaluate the importance of each feature, and the data is divided into different subspaces according to the importance of each feature, and the corresponding weight is assigned to each subspace. Then, the isolation forest algorithm is applied in each subspace for anomaly detection, and the anomaly score of each subspace is obtained. Finally, the anomaly score and weight of each subspace were combined to obtain the final anomaly detection score. In order to evaluate the effectiveness of the algorithm, the proposed algorithm was compared with other four algorithms on the credit loan fraud data set. The results show that the AUC index, accuracy, recall rate and F1 score of the proposed algorithm are higher than those of the comparison algorithms, showing high effectiveness.

Integrating proximal geophysical sensing and machine learning for digital soil mapping: Spatial prediction and model evaluation using a small dataset

Geophysical methods support soil security by providing non-invasive tools to assess soil properties, monitor degradation, and guide sustainable management strategies. However, studies focusing the spatial prediction of geophysical data remain limited. In this research, we aimed to model and predict the spatial distribution of soil geophysical properties using parent material and terrain attributes with machine learning algorithms. In addition, we tested the nested leave-one-out cross validation (nested-LOOCV) method to deal with datasets with limited size. We performed a geophysical survey using three types of sensors (radiometric, magnetic and electric methods). The random forest (RF) and support vector machine (SVM) algorithms presented the best results, with RF showing higher performance for K40 and magnetic susceptibility, and SVM had higher performance for eU, eTh and apparent electrical conductivity. Parent materials and digital elevation model were the most significant variables for the modelling. The nested-LOOCV method proved to be adequate for small soil dataset. Machine learning techniques are potential tools for modelling soil geophysical variables. The combination with computational techniques shows the great relevance of geophysical measurements for the estimation of soil properties related to fertility and soil genesis.

Satellites Reveal Global Migration Patterns of Natural Mountain Treelines during Periods of Rapid Warming

Profound global transformations in the Anthropocene epoch are hastening shifts in species ranges, with natural mountain treeline migration playing a crucial role in this overarching species movement. The varied reactions of mountain treelines to climatic conditions across diverse climatic zones, when compounded by local disturbances, result in distinct migration patterns. Usually, warming encourages mountain treelines to migrate to higher elevations. Nevertheless, in a period of rapid warming, it remains unclear whether the natural mountain treeline in global thermal climatic zones and subclimatic zones has expedited its upward movement. Here, we employed remote sensing observations and the random forest algorithm to investigate the natural treeline dynamics across 24 major mountain ranges worldwide amidst a period of rapid warming (1990–2020). Our research shows substantial disparities in the migration patterns of natural mountain treelines across the global thermal zone. The natural mountain treeline in tropical and subtropical zones descends by an average of 1.1 and 0.8 m per year, respectively. Only 18.8 and 35.5% of the natural mountain treelines in these regions had undergone upward migration, respectively. The average migration rates of natural mountain treelines in temperate and boreal zones were 0.7 m per year. Correspondingly, 47 and 33.2% of the natural mountain treelines in these zones had already shifted to higher elevations. The highest average migration rate of natural mountain treelines occurs in temperate continental climates (1.7 m per year). The loss or degradation of alpine species habitats, a direct consequence of the upward movement of the treeline, highlights the necessity for increased monitoring and protection of alpine species in temperate and boreal zones in the future.

Analysis and prediction of energy consumption in office buildings with variable refrigerant flow systems: A case study

Accurately predicting building air-conditioning energy consumption is particularly important for energy management. However, for small datasets, there is a lack of simple and effective prediction methods applicable to multiple buildings. In order to fully explore the energy consumption mode of Variable Refrigerant Flow (VRF) air conditioning system and improve the accuracy of the model. This study uses the weather data of an outdoor meteorological data collection system and the air conditioning system and energy consumption data of an office building in Jinan. Initial analysis revealed that occupant behavior and daily variations in outdoor air temperatures significantly impact the energy consumption of air conditioning systems. Then, by applying the random forest algorithm and Pearson correlation analysis, the study identified daily average outdoor air temperature and VRF indoor unit running time as critical features for prediction. Subsequently, three machine learning models—Multiple Linear Regression (MLR), Back Propagation Neural Network (BPNN), and Long Short-Term Memory (LSTM) neural network—were developed and assessed using a 4:1 training-to-testing set ratio. Comparative analysis revealed that all models performed robustly, with the MLR model exhibited superior predictive accuracy (R2 (Coefficient of Determination) = 0.98, MAE (Mean Absolute Error) = 0.005 kWh/(m2·day), and RMSE (Root Mean Square Error) = 0.007 kWh/(m2·day) for the testing set) and simplicity. Furthermore, the effectiveness of the MLR model in handling small sample sizes was confirmed through cross-validation of data from two additional office buildings, highlighting its suitability for predicting energy consumption of VRF air conditioning systems in office buildings.

Avian diversity across guilds in North America versus vegetation structure as measured by the Global Ecosystem Dynamics Investigation (GEDI)

Avian diversity, a key indicator of ecosystem health, is closely related to canopy structure. Most avian diversity models are based on either optical remote sensing or airborne lidar data, but the latter is limited to small study areas. The launch of the Global Ecosystem Dynamics Investigation (GEDI) instrument in 2018 has opened new avenues for exploring the influence of vegetation structure on avian diversity. To examine how direct measurements of canopy structural characteristics explain bird diversity across North America, we analyzed 18 GEDI metrics from 2019 to 2022, along with corresponding Breeding Bird Survey (BBS) counts and AVONET morphological data, analyzing effects across broad regions and at varying spatial extents. We grouped 440 bird species into 20 ecological guilds under six guild categories and employed random forest algorithms to model avian diversity across eight spatial extents (1, 2, 3, 4, 5, 10, 20, and 39.2 km). The models predicted six diversity indices, including species richness (sRich), functional richness (fRich), evenness (fEve), dispersion (fDis), divergence (fDiv), and redundancy (fRed) across eight spatial extents. The best-predicted guilds varied for each diversity index. The most accurate models were sRich (pseudo-R2 = 0.71, RMSE = 4.28) and fRed (pseudo-R2 = 0.60, RMSE = 0.13) for forest specialists guilds; fRich (pseudo-R2 = 0.55, RMSE = 0.18) for urban guilds; fEve (pseudo-R2 = 0.28, RMSE = 0.08) for insectivore guilds; and fDiv (pseudo-R2 = 0.38, RMSE = 0.12) and fDis (pseudo-R2 = 0.53, RMSE = 0.87) for short distance migrants guilds. Our results highlight the critical role of canopy structure, including its horizontal and vertical distribution and variation, in predicting avian diversity, as measured by the mean number of detected modes (num_detectedmodes), the standard deviation of foliage height diversity (FHD), num_detectedmodes, canopy cover, and plant area index (PAI) across the spatial extents centered on BBS routes. Therefore, we recommend incorporating the GEDI metrics into avian diversity modeling and mapping across North America, thereby potentially enhancing bird habitat management and conservation efforts.

Regional Logistics Demand Prediction Using BP Neural Network Based on Particle Swarm Optimization

Accurate regional logistics demand prediction is an indispensable part of a scientific and reasonable logistics construction system. This article mainly uses machine learning methods to screen features, select effective features, and then uses intelligent algorithms to optimize the BP neural network. The optimized model is trained and predicted for the future, optimizing research on regional logistics demand prediction, and providing effective suggestions for the development of the logistics industry. This article first uses the random forest algorithm to select features, eliminate redundant features, and then uses particle swarm optimization algorithm to optimize the parameters of the BP neural network to improve model performance. Finally, grey prediction is used to construct a prediction dataset, and the future trend of regional logistics demand is predicted, based on which suggestions are proposed. The results show that the optimized combination model proposed in this article can effectively improve prediction accuracy and make an effective contribution to the development of regional logistics.

Super Learner Algorithm for Carotid Artery Disease Diagnosis: A Machine Learning Approach Leveraging Craniocervical CT Angiography.

This study introduces a machine learning (ML) approach to diagnosing carotid artery diseases, including stenosis, aneurysm, and dissection, by leveraging craniocervical computed tomography angiography (CTA) data. A meticulously curated, balanced dataset of 122 patient cases was used, ensuring reproducibility and data quality, and this is publicly accessible at (insert dataset location). The proposed method integrates a super learner model which combines adaptive boosting, gradient boosting, and random forests algorithms, achieving an accuracy of 90%. To enhance model robustness and generalization, techniques such as k-fold cross-validation, bootstrapping, data augmentation, and the synthetic minority oversampling technique (SMOTE) were applied, expanding the dataset to 1000 instances and significantly improving performance for minority classes like aneurysm and dissection. The results highlight the pivotal role of blood vessel structural analysis in diagnosing carotid artery diseases and demonstrate the superior performance of the super learner model in comparison with state-of-the-art (SOTA) methods in terms of both accuracy and robustness. This manuscript outlines the methodology, compares the results with state-of-the-art approaches, and provides insights for future research directions in applying machine learning to medical diagnostics.

Proposing Optimized Random Forest Models for Predicting Compressive Strength of Geopolymer Composites

This paper explores advanced machine learning approaches to enhance the prediction accuracy of compressive strength (CoS) in geopolymer composites (GePC). Geopolymers, as sustainable alternatives to Ordinary Portland Cement (OPC), offer significant environmental benefits by utilizing industrial by-products such as fly ash and ground granulated blast furnace slag (GGBS). The accurate prediction of their compressive strength is crucial for optimizing their mix design and reducing experimental efforts. We present a comparative analysis of two hybrid models, Harris Hawks Optimization with Random Forest (HHO-RF) and Sine Cosine Algorithm with Random Forest (SCA-RF), against traditional regression methods and classical models like the Extreme Learning Machine (ELM), General Regression Neural Network (GRNN), and Radial Basis Function (RBF). Using a comprehensive dataset derived from various scientific publications, we focus on key input variables including the fine aggregate, GGBS, fly ash, sodium hydroxide (NaOH) molarity, and others. Our results indicate that the SCA-RF model achieved a superior performance with a root mean square error (RMSE) of 1.562 and a coefficient of determination (R2) of 0.987, compared to the HHO-RF model, which obtained an RMSE of 1.742 and an R2 of 0.982. Both hybrid models significantly outperformed traditional methods, demonstrating their higher accuracy and reliability in predicting the compressive strength of GePC. This research underscores the potential of hybrid machine learning models in advancing sustainable construction materials through precise predictive modeling, paving the way for more environmentally friendly and efficient construction practices.

Predicting Students' Academic Performances Using Machine Learning Algorithms in Educational Data Mining

In every culture and era, education is considered the most fundamental reality and rule that societies prioritize and deem essential. Throughout the process spanning thousands of years, from the emergence of writing to the present day, education has undergone various forms and formats of change. Education has been a continuous guide for shaping, influencing, sustaining societies, and maintaining its dynamics throughout these historical processes. The continuous evolution and growth of education systems and formats worldwide, with changes affecting the quality of education, have the potential to influence nations and societies in every field, ultimately leading to the emergence of an informed society, achievable only through quality education. In this study, the aim is to determine the factors affecting students' academic performance and predict students' end-of-term academic grades using machine learning algorithms within the scope of Earned Value Management (EVM). Such studies have great potential to increase efficiency in education, improve student achievement and improve education policies. With the use of machine learning algorithms, these goals can be achieved more quickly and efficiently. Five different machine learning algorithms, namely RF, KA, KNN, SVM, and NB, have been employed in the study. Binary and multiclass classification methods were used in prediction processes, and among these methods, the Random Forest (RF) algorithm achieved the highest success prediction rates of 0.97 and 0.93, respectively, in both classification methods.

Prediction of adverse drug reactions using demographic and non-clinical drug characteristics in FAERS data

The presence of adverse drug reactions (ADRs) is an ongoing public health concern. While traditional methods to discover ADRs are very costly and limited, it is prudent to predict ADRs through non-invasive methods such as machine learning based on existing data. Although various studies exist regarding ADR prediction using non-clinical data, a process that leverages both demographic and non-clinical data for ADR prediction is missing. In addition, the importance of individual features in ADR prediction has yet to be fully explored. This study aims to develop an ADR prediction model based on demographic and non-clinical data, where we identify the highest contributing factors. We focus our efforts on 30 common and severe ADRs reported to the Food and Drug Administration (FDA) between 2012 and 2023. We have developed a random forest (RF) and deep learning (DL) machine learning model that ingests demographic data (e.g., Age and Gender of patients) and non-clinical data, which includes chemical, molecular, and biological drug characteristics. We successfully unified both demographic and non-clinical data sources within a complete dataset regarding ADR prediction. Model performances were assessed via the area under the receiver operating characteristic curve (AUC) and the mean average precision (MAP). We demonstrated that our parsimonious models, which include only the top 20 most important features comprising 5 demographic features and 15 non-clinical features (13 molecular and 2 biological), achieve ADR prediction performance comparable to a less practical, feature-rich model consisting of all 2,315 features. Specifically, our models achieved an AUC of 0.611 and 0.674 for RF and DL algorithms, respectively. We hope our research provides researchers and clinicians with valuable insights and facilitates future research designs by identifying top ADR predictors (including demographic information) and practical parsimonious models.

Comprehensive Analysis of Bile Medicines Based on UHPLC-QTOF-MSE and Machine Learning.

Based on UHPLC-QTOF-MSE analysis and quantized processing, combined with machine learning algorithms, data modeling was carried out to realize digital identification of bear bile powder (BBP), chicken bile powder (CIBP), duck bile powder (DBP), cow bile powder (CBP), sheep bile powder (SBP), pig bile powder (PBP), snake bile powder (SNBP), rabbit bile powder (RBP), and goose bile powder (GBP). First, 173 batches of bile samples were analyzed by UHPLC-QTOF-MSE to obtain the retention time-exact mass (RTEM) data pair to identify bile acid-like chemical components. Then, the data were modeled by combining support vector machine (SVM), random forest (RF), artificial neural network (ANN), gradient boosting (GB), AdaBoost (AB), and Naive Bayes (NB), and the models were evaluated by the parameters of accuracy (Acc), precision (P), and area under the curve (AUC). Finally, the bile medicines were digitally identified based on the optimal model. The results showed that the RF model constructed based on the identified 12 bile acid-like chemical constituents and random forest algorithm is optimal with ACC, P, and AUC > 0.950. In addition, the accuracy of external identification verification of 42 batches of bile medicines detected at different times is 100.0%. So based on UHPLC-QTOF-MSE analysis and combined with the RF algorithm, it can efficiently and accurately realize the digital identification of bile medicines, which can provide reference and assistance for the quality control of bile medicines. In addition, hyodeoxycholic acid, glycohyodeoxycholic acid, and taurochenodeoxycholic acid, and so forth are the most important bile acid constituents for the identification of nine bile medicines.

Effects of Climate Changes and Crop Phenological Responses on Soil Organic Carbon of Cultivated Land in Fujian Province

Research on the mechanism of how climate change affects cultivated soil organic carbon is the basis for the management of cultivated land quality in the context of climate change. Crop phenological responses to climate change have an important effect on cultivated soil organic carbon as well. However, previous research primarily focused on the independent effects of climate change or crop phenological responses on the changes in soil organic carbon, and few studies have analyzed the changes in cultivated soil organic carbon under the combined influence of both factors or quantified their contribution rates to the changes in cultivated soil organic carbon. Based on topsoil samples in 2008 and 2021, annual pre-season and mid-season climate data from 2008 to 2021, and the phenological parameters extracted from the enhanced vegetation index （EVI） time series from 2007 to 2022, a soil organic carbon predictive model was constructed using the random forest algorithm. The total change in soil organic carbon from 2008 to 2021, the change in soil organic carbon under climate change alone, and the change in soil organic carbon under the synergistic influence of climate change and crop phenological responses were simulated. Furthermore, the contributions of climate change and crop phenological responses to the changes in cultivated soil organic carbon were distinguished and quantified. Moreover, the dominant influencing factors of soil organic carbon changes and their spatial distributions were identified and analyzed. The results were as follows： ① Under the synergistic influence of climate change and crop phenological responses, a decrease was observed in soil organic carbon in 74.15% of the cultivated land area in Fujian Province during the years 2008-2021, with an average decrease of 2.20 g·kg-1. Additionally, there was an increase in soil organic carbon in 25.85% of the cultivated area, with an average increase of 1.48 g·kg-1. ②The average contribution rates of pre-season climate, crop phenological responses to climate change, mid-season climate, and phenological changes resulting from cultivars shifts or other adjustments of agricultural measures to soil organic carbon changes were 34.08%, 28.56%, 22.75%, and 14.61%, respectively. Overall, climate change had a greater impact on the changes in cultivated soil organic carbon in Fujian Province than the crop phenological response to climate change. ③ The regions where climate change and phenological response jointly acted as dominant influencing factors held the largest area, accounting for 47.06% of the total cultivated land area in Fujian Province, and the regions where climate change was the dominant influencing factor alone held the second-largest area, accounting for 28.64% of the total cultivated land area. ④ Higher contribution rates of pre-season climate factors and phenological changes resulting from cultivar shifts or other adjustments of agricultural measures tended to be distributed in higher-altitude areas, whereas higher contribution rates of mid-season climate factors and phenological responses to climate change tended to be distributed in lower-altitude areas. These research findings can provide a theoretical basis for decision making regarding the management of cultivated land quality and the safeguarding of food security in the context of climate change.