Article
Forestry
Chen Chang, Zhongke Feng, Ziye Liu
Summary: With the optimization of random forest parameters, this study explores the impact of environmental factors on tree density. The results indicate that average temperature, soil thickness, and forest water consumption are the main factors limiting tree density, and the influence of each factor varies depending on the stage of tree growth. Based on forest resource data, tree density distribution grid maps were generated using models and interpolation methods, providing theoretical and data support for the development of appropriate forest management strategies.
Article
Medicine, Research & Experimental
Mahdi Akbarzadeh, Nadia Alipour, Hamed Moheimani, Asieh Sadat Zahedi, Firoozeh Hosseini-Esfahani, Hossein Lanjanian, Fereidoun Azizi, Maryam S. Daneshpour
Summary: This study compared different machine learning classification methods in predicting the status of metabolic syndrome and identifying influential factors. The findings showed that machine learning models outperformed conventional statistical approaches and can be used to identify individuals at high risk of developing metabolic syndrome.
JOURNAL OF TRANSLATIONAL MEDICINE
(2022)
Article
Environmental Sciences
Wenjie Hou, Guanghua Yin, Jian Gu, Ningning Ma
Summary: In this study, a hybrid RF-SVR-PSO model was used to estimate the daily evapotranspiration of spring maize. The results showed that this model performed better than standalone models for ETc estimation in semi-arid regions.
Article
Biochemistry & Molecular Biology
Ahmad Firoz, Adeel Malik, Hani Mohammed Ali, Yusuf Akhter, Balachandran Manavalan, Chang-Bae Kim
Summary: In this study, a new two-layer hybrid framework called PRR-HyPred was constructed to simultaneously predict and classify PRRs. Using support vector machine and random forest-based classifier, PRR-HyPred achieved accuracies of 83.4% and 95% in the first and second layers respectively. This is the first study that can predict and classify PRRs into specific families, and it can be a valuable tool for large-scale PRR prediction and classification, facilitating future studies.
INTERNATIONAL JOURNAL OF BIOLOGICAL MACROMOLECULES
(2023)
Article
Biodiversity Conservation
Xi Wu, Wenjiao Shi, Fulu Tao
Summary: Hydrological models and remote sensing methods are often used to estimate water retention amounts and analyze spatial variations. However, their accuracies are lower than site-observational data. This study used observational sites and a RF model to predict canopy interception, litter water-holding amount, soil water storage, and forest water retention in China, finding significant spatial variations in different forest types across different basins.
ECOLOGICAL INDICATORS
(2021)
Article
Computer Science, Artificial Intelligence
Arwinder Dhillon, Ashima Singh, Vinod Kumar Bhalla
Summary: This research proposes a framework called BioSurv for identifying cancer biomarkers and predicting cancer survival using machine learning and deep learning techniques. Multi-omics data from breast cancer and lung adenocarcinoma are analyzed, and statistical tests and an optimization algorithm are employed for feature selection. Thirteen BRCA and fifteen LUAD poor prognostic markers are identified, and a Bayesian optimized deep neural network achieves high accuracy in cancer survival prediction for both types of cancer.
APPLIED SOFT COMPUTING
(2023)
Article
Ecology
Robin Singh Bhadoria, Manish Kumar Pandey, Pradeep Kundu
Summary: Human intervention causing forest fires hinders nature's ability to recover, leading to climate change consequences that we must take responsibility for and minimize. Mitigating fires by predicting and controlling their spread can be enhanced through machine learning models, like the proposed RVFR model, which achieves higher accuracy in predicting forest fires based on past data.
ECOLOGICAL INFORMATICS
(2021)
Article
Environmental Sciences
Chunfang Kong, Yiping Tian, Xiaogang Ma, Zhengping Weng, Zhiting Zhang, Kai Xu
Summary: This study used different models to evaluate landslide susceptibility in Zhaoping County, and found that the PSO-RF model had the highest accuracy. The PSO algorithm had a good effect on the SVM and RF models, and all four models performed well for landslide susceptibility evaluation.
Article
Environmental Studies
Xiaochen Liu, Zhenxing Bian, Zhentao Sun, Chuqiao Wang, Zhiquan Sun, Shuang Wang, Guoli Wang
Summary: This study verifies the feasibility of integrating landscape patterns in predicting soil organic carbon (SOC) in the Lower Liaohe Plain and finds that landscape variables improve mapping accuracy compared to natural variables. By considering the overall climate change, topographical changes, and human activities, landscape patterns can better characterize the impacts of human activities on SOC.
Article
Agriculture, Dairy & Animal Science
Cem Tirink, Dariusz Piwczynski, Magdalena Kolenda, Hasan Onder
Summary: This study aimed to estimate body weight from various biometric measurements and features using data mining and machine learning algorithms. The results showed that the random forest algorithm can help improve important characteristics and breed an elite population in Poland.
Article
Multidisciplinary Sciences
Friederike Maite Siemers, Juergen Bajorath
Summary: This paper explores the application of random forest (RF) and support vector machine (SVM) in molecular machine learning (ML) and compound property prediction. By utilizing explainable artificial intelligence (XAI) methods, such as the Shapley value concept, the study reveals that RF and SVM models have different learning characteristics in their predictions and chemically intuitive explanations for accurate predictions originate from different sources.
SCIENTIFIC REPORTS
(2023)
Article
Engineering, Electrical & Electronic
Guo-Feng Fan, Yan-Rong Liu, Hui-Zhen Wei, Meng Yu, Yin-He Li
Summary: This paper proposes a hybrid model based on EEMD-RF-SVR-RR algorithm for electric load forecasting. Numerical experiments have shown that the model outperforms other models in terms of forecasting accuracy, confirming its feasibility and effectiveness in short-term load forecasting.
ELECTRIC POWER SYSTEMS RESEARCH
(2022)
Article
Chemistry, Analytical
Michelle Smit, Seer J. Ikurior, Rene A. Corner-Thomas, Christopher J. Andrews, Ina Draganova, David G. Thomas
Summary: Animal behaviour can be monitored using triaxial accelerometers to identify specific behaviours, providing a non-invasive and continuous method of observation. Machine learning models, such as random forest and supervised self-organizing map, can be used to predict behaviours based on accelerometer data. The results of this study show that triaxial accelerometers are capable of accurately identifying cat-specific behaviours.
Article
Pharmacology & Pharmacy
Xiaoda Yang, Hongshun Qiu, Yuxiang Zhang, Peijian Zhang
Summary: The study aims to predict the inhibitory effect of amide derivatives on xanthine oxidase (XO) by building several models based on the theory of quantitative structure-activity relationship (QSAR). Linear and non-linear models were constructed using the heuristic method (HM) and XGBoost, respectively. Among the non-linear models, MIX-SVR method achieved the best result by combining different kernel functions.
FRONTIERS IN PHARMACOLOGY
(2023)
Article
Environmental Sciences
Tao Zhou, Yuting Hou, Zhihan Yang, Benjamin Laffitte, Ke Luo, Xinrui Luo, Dan Liao, Xiaolu Tang
Summary: This study used a random forest model to predict the variation of global net primary production (NPP) at different spatial resolutions. The results showed that NPP exhibited similar spatial patterns and interannual variation trends at different resolutions, but there were significant differences in total global NPP. Therefore, the study emphasizes the importance of selecting an appropriate spatial resolution when modeling carbon fluxes.
SCIENCE OF THE TOTAL ENVIRONMENT
(2023)
Article
Computer Science, Artificial Intelligence
Guiliang Gong, Jiuqiang Tang, Dan Huang, Qiang Luo, Kaikai Zhu, Ningtao Peng
Summary: This paper proposes a flexible job shop scheduling problem with discrete operation sequence flexibility and designs an improved memetic algorithm to solve it. Experimental results show that the algorithm outperforms other algorithms in terms of performance. The proposed model and algorithm can help production managers obtain optimal scheduling schemes considering operations with or without sequence constraints.
SWARM AND EVOLUTIONARY COMPUTATION
(2024)
Article
Computer Science, Artificial Intelligence
Daniel Molina-Perez, Efren Mezura-Montes, Edgar Alfredo Portilla-Flores, Eduardo Vega-Alvarado, Barbara Calva-Yanez
Summary: This paper presents a new proposal based on two fundamental strategies to improve the performance of the differential evolution algorithm when solving MINLP problems. The proposal considers a set of good fitness-infeasible solutions to explore promising regions and introduces a composite trial vector generation method to enhance combinatorial exploration and convergence capacity.
SWARM AND EVOLUTIONARY COMPUTATION
(2024)