Random Forest Machine Learning Algorithm Research Articles

Development of Random Forest Model for Stroke Prediction

Stroke is a significant cause of mortality and morbidity worldwide, and early detection and prevention of stroke are essential for improving patient outcomes. Machine learning algorithms have been used in recent years to predict the risk of stroke by leveraging large amounts of clinical and demographic data. The development of a stroke prediction system using Random Forest machine learning algorithm is the main objective of this thesis. The primary goal of the project is to increase the accuracy of stroke detection while addressing the shortcomings of the current system, which include real- time deployment and interpretability issues with logistic regression. The development and use of an ensemble machine learning-based stroke prediction system, performance optimization through the use of ensemble machine learning algorithms, performance assessment, and real-time model deployment through the use of Python Django are among the goals of the research. The study's potential to improve public health by lessening the severity and consequences of strokes through early diagnosis and treatment makes it significant. Data collection, preprocessing, model selection, evaluation, and real-time deployment using Python Django are all part of the research technique. Our dataset consists of 5110 rows of tuples and columns with total size of 69kg. The performance of our stroke prediction algorithm was evaluated using confusion metrics-consisting of accuracy, precision, recall and F1-score. At the end of the research, Random Forest model gave an accuracy of 98.5% compared to the existing model logistic regression which has 86% accuracy.

Construction of diagnostic models for the progression of hepatocellular carcinoma using machine learning.

Liver cancer is one of the most prevalent forms of cancer worldwide. A significant proportion of patients with hepatocellular carcinoma (HCC) are diagnosed at advanced stages, leading to unfavorable treatment outcomes. Generally, the development of HCC occurs in distinct stages. However, the diagnostic and intervention markers for each stage remain unclear. Therefore, there is an urgent need to explore precise grading methods for HCC. Machine learning has emerged as an effective technique for studying precise tumor diagnosis. In this research, we employed random forest and LightGBM machine learning algorithms for the first time to construct diagnostic models for HCC at various stages of progression. We categorized 118 samples from GSE114564 into three groups: normal liver, precancerous lesion (including chronic hepatitis, liver cirrhosis, dysplastic nodule), and HCC (including early stage HCC and advanced HCC). The LightGBM model exhibited outstanding performance (accuracy = 0.96, precision = 0.96, recall = 0.96, F1-score = 0.95). Similarly, the random forest model also demonstrated good performance (accuracy = 0.83, precision = 0.83, recall = 0.83, F1-score = 0.83). When the progression of HCC was categorized into the most refined six stages: normal liver, chronic hepatitis, liver cirrhosis, dysplastic nodule, early stage HCC, and advanced HCC, the diagnostic model still exhibited high efficacy. Among them, the LightGBM model exhibited good performance (accuracy = 0.71, precision = 0.71, recall = 0.71, F1-score = 0.72). Also, performance of the LightGBM model was superior to that of the random forest model. Overall, we have constructed a diagnostic model for the progression of HCC and identified potential diagnostic characteristic gene for the progression of HCC.

Utilizing Machine Learning for Context-Aware Digital Biomarker of Stress in Older Adults

Identifying stress in older adults is a crucial field of research in health and well-being. This allows us to take timely preventive measures that can help save lives. That is why a nonobtrusive way of accurate and precise stress detection is necessary. Researchers have proposed many statistical measurements to associate stress with sensor readings from digital biomarkers. With the recent progress of Artificial Intelligence in the healthcare domain, the application of machine learning is showing promising results in stress detection. Still, the viability of machine learning for digital biomarkers of stress is under-explored. In this work, we first investigate the performance of a supervised machine learning algorithm (Random Forest) with manual feature engineering for stress detection with contextual information. The concentration of salivary cortisol was used as the golden standard here. Our framework categorizes stress into No Stress, Low Stress, and High Stress by analyzing digital biomarkers gathered from wearable sensors. We also provide a thorough knowledge of stress in older adults by combining physiological data obtained from wearable sensors with contextual clues from a stress protocol. Our context-aware machine learning model, using sensor fusion, achieved a macroaverage F-1 score of 0.937 and an accuracy of 92.48% in identifying three stress levels. We further extend our work to get rid of the burden of manual feature engineering. We explore Convolutional Neural Network (CNN)-based feature encoder and cortisol biomarkers to detect stress using contextual information. We provide an in-depth look at the CNN-based feature encoder, which effectively separates useful features from physiological inputs. Both of our proposed frameworks, i.e., Random Forest with engineered features and a Fully Connected Network with CNN-based features validate that the integration of digital biomarkers of stress can provide more insight into the stress response even without any self-reporting or caregiver labels. Our method with sensor fusion shows an accuracy and F-1 score of 83.7797% and 0.7552, respectively, without context and 96.7525% accuracy and 0.9745 F-1 score with context, which also constitutes a 4% increase in accuracy and a 0.04 increase in F-1 score from RF.

Urban lakes change extraction using time series GaoFen-1 satellite imagery

Abstract. Urban lakes serve an indispensable role in maintaining the ecological balance of cities, ensuring flood safety, and providing recreational spaces for tourism. With the development of human activities and economic, the extent of urban lakes are inevitably influenced. Currently, the ability to detect detailed temporal changes in urban lake areas using high resolution data still has limitations. This study proposed a novel method by combining time series Gaofen-1 (GF-1) remote sensing data and random forest machine learning algorithm to explore the urban lakes change Zhushan Lake located in Wuhan. The research conducted the extraction of surface water for Zhushan Lake and its surrounding pit-ponds from 2013 to 2020. And then, a quantitative analysis of the characteristics and driving factors of lake changes is conducted. We find that (1) the accuracy of surface water extraction using the random forest classification method consistently exceeded 96%. The Kappa coefficient ranges from a minimum of 0.86 to a maximum of 0.99. (2) A noticeable decline was observed in the water areas of Zhushan Lake and its surrounding pit-ponds, predominantly along the northwestern shoreline and in the eastern pond regions. This decline is primarily attributed to pressures from building construction. The methodology proposed in this study is suitable for the area management of lakes in urban areas.

Analysis and Forecasting of Energy Data and GDP Based on Pearson Correlation Analysis-Random Forest Algorithm

The relationship between energy data and GDP is an important economic issue because energy is the foundation that supports modern economic development. This paper examines the relationship between energy data and GDP and uses Pearson correlation analysis to calculate the relationship between each variable in energy data and annual GDP for 52 regions. The results show that total electricity consumption is the factor with the highest degree of influence on GDP, followed by natural gas consumption, biomass consumption, hydroelectricity consumption, LPG consumption and geothermal energy consumption. This indicates that there is an extremely strong correlation between local energy data and GDP. To further explore this correlation, this paper attempts to predict the GDP of the locality based on the local energy data using a machine learning approach.The Random Forest algorithm is chosen as the machine learning model, and the training, validation and test sets are divided according to the ratio of 6:2:2. By using MSE, RMSE, MAE, MAPE and R to evaluate the model, it is found that the random forest machine learning algorithm can predict the local GDP well.The predicted value of GDP and the actual value of GDP are very close to each other, meanwhile, the value of MSE is 62.779.These findings show that the local energy data are closely related to economic development. The impact of energy factors on the economy should be taken into account when formulating economic policies. In addition, machine learning algorithms can provide strong support for economic forecasting and provide more accurate information for decision makers.

Biomass energy transformation: Harnessing the power of explainable ai to unlock the potential of ultimate analysis data

Biomass fuels are a promising renewable energy source for both heat and electricity generation. When weighing the economics of biomass energy, the higher heating value (HHV) is a significant factor to consider. To estimate HHVs, a full understanding of the chemical constituents of biomass is required, which can be costly and time-consuming to find experimentally. To solve this issue, this study applies three modern and robust machine learning algorithms XGBoost, Random Forest (RF), and Adaptive Boosting (AdaBoost), to predict HHVs reliably based on final analysis data, which includes carbon (C, %), hydrogen (H, %), nitrogen (N, %), and oxygen (O, %) content in biomass. The linear regression (LR) was used for measuring baseline performance. The XGBoost, AdaBoost, and RF models exhibit endurance in model prediction, however, the XGBoost model outperforms all other models. A Bayesian strategy for hyperparameter optimization was employed to optimize the prediction outcomes. The R2 values achieved during the model testing for LR, RF, XGBoost, and AdaBoost-based models were 0.8425, 0.9822, 0.9967, and 0.8367, respectively. Once the best-performing model (XGBoost) was determined, it was evaluated with Shapley additive explanations (SHAP) for feature analysis. This study report adds to informed decision-making and instills trust in biomass energy applications by putting light on the fundamental mechanisms driving HHVs.

Soil moisture retrieval at high spatial resolution over alpine ecosystems on Nagqu-Tibetan plateau: A comparative study on semiempirical and machine learning approaches

Soil moisture (SM) is an essential climate variable that directly and indirectly affects vegetation growth and survival through land‒atmosphere interactions. Alpine vegetation on the Tibetan Plateau is part of a unique ecosystem that is vulnerable to changes in environmental factors such as SM; consequently, this makes this ecosystem extremely sensitive to climate change. This study investigated the potential of synthetic aperture radar (SAR) vegetation indices based on Sentinel-1 data for retrieving SM at high spatial resolution (10 m) over an alpine grassland ecosystem in the Nagqu region. Several SAR vegetation indices, including the dual polarization SAR vegetation index (DPSVI), modified dual polarization SAR vegetation index (mDPSVI), dual polarimetric radar vegetation index (DpRVI), polarimetric radar vegetation index (PRVI), and radar vegetation index (RVI), were used in the semiempirical water cloud model (WCM) to determine which indices provide better SM retrievals in this alpine ecosystem. In addition, the potential of the distributed random forest (DRF) machine learning algorithm was explored using the same variables as the WCM together with several ecohydrological parameters from different data sources. The recursive feature elimination algorithm was used to establish the optimized DRF model. Among the vegetation indices based on SAR data, DPSVI, DpRVI, and PRVI showed similar results, with DPSVI performing slightly better than the other SAR indices, with a correlation coefficient (R2) of 0.70 and root mean squared error (RMSE) of 0.04 m3m-3. A comparison of the optimized DRF with the best fitted WCM reveals that the DRF algorithm outperformed the WCM, including having more predictors (10 variables) in the model. The results show that the overall accuracies in terms of the R2 values and the RMSEs of both the WCMs and the DRF models were 0.52–0.75 and 0.08 m3 m−3 to 0.04 m3 m−3, respectively, which was validated over in situ SM measurements in the Nagqu region.

A Comparison of Pixel Based and Object Based Image Classification for Cropland Area Estimation

Pixel-based (PB) and object-based (OB) Land Use–Land Cover (LULC) classification techniques can be applied in Google Earth Engine (GEE), a flexible cloud-based platform, because of its numerous state-of-the-art functions that comprise several Machine Learning (ML) methods. Adding some texture measure, any measure to a classification usually improves the accuracy. Object segmentation and object textural analysis are two OB methods that are still uncommon in the GEE environment. Object based image classification is difficult because it can be challenging to concatenate the correct functions and adjust various parameters in order to get past the computational limitations of GEE. The goal of this work is to develop and test an OB classification approach that combines the ML algorithm Random Forest (RF) to perform the final classification, the Gray-Level Co-occurrence Matrix (GLCM) to calculate cluster textural indices and the Simple Non-Iterative Clustering (SNIC) algorithm to identify spatial clusters. The primary seven GLCM indices are subjected to a Principal Components Analysis (PCA) in order to combine the textural data needed for the OB classification into a single band. The proposed methodology was broadly tested in a 304 km2 study area, located in the Telangana state (India), using Sentinel 2 (S2).

Machine learning-based potential loss assessment of maize and rice production due to flash flood in Himachal Pradesh, India.

Flash floods in mountainous regions like the Himalayas are considered to be common natural calamities. Their consequences often are more dangerous than any flood event in the plains. These hazards not only put human lives at threat but also cause economic deflation due to the loss of lands, properties, and agricultural production. Hence, assessing the impact of such hazards in the existing agricultural system is of utmost importance to understand the probable crop loss. In this paper, we studied the efficiency of the remotely sensed microwave data to map the croplands affected by the flash flood that occurred in July 2023 in Himachal Pradesh, a mountainous state in the Indian Himalayan Region. The Una, Hamirpur, Kangra, and Sirmaur districts were identified as the most affected areas, with about 9%, 6%, 5.74%, and 3.61% of the respective districts' total geographical area under flood. Further, four machine learning algorithms (random forest, support vector regressor, k-nearest neighbor, and extreme gradient boosting) were evaluated to forecast maize and rice crop production and potential loss during the Kharif season in 2023. A regression algorithm with ten predictor variables consisting of the cropland area, two vegetation indices, and seven climatic parameters was applied to forecast the maize and rice production in the state. Amongst the four algorithms, random forest showed outstanding performance compared to others. The random forest regressor estimated the production of maize and rice with R2 more than 0.8 in most districts. The mean absolute error and the root mean squared error obtained from the random forest regressor were also minimal compared to the others. The maximum production loss of maize is estimated for Solan (54.13%), followed by Una (11.06%), and of rice in Kangra (19.1%), Una (18.8%) and Kinnaur (18.5%) districts. This indicated the utility of the proposed approach for a quick in-season forecast on crop production loss due to climatic hazards.

Geotechnologies on the phosphorus stocks determination in tropical soils: General impacts on society

Phosphorus (P) is a critical nutrient for primary production in terrestrial and aquatic ecosystems. As P mineral reserves are finite and non-renewable, there is an increasing discussion on its sustainable utilization to safeguard food security for future generations. Understanding the spatial distribution of soil P is central in advancing effective phosphorus management and fostering sustainable agricultural practices. This study aims to digitally map the stocks of available P (AP) and total P (TP) in Brazil at a fine resolution (30 m). Using the Random Forest machine learning algorithm and a database of topsoil (0-20 cm) with 28,572 samples for AP and 3154 for TP, we predicted P stocks based on environmental covariates related to soil formation processes. By dividing Brazil into two sub-regions, representing areas with native coverage and anthropogenic ones, we built independent predictive models for each sub-region. Our results show that Brazil has a TP stock of 531 Tg and an AP stock of 17.4 Tg. The largest soil TP stocks are in the Atlantic Forest biome (73.8 g.m2), likely due to higher organic carbon stocks in this biome. The largest AP stocks were in the Caatinga biome (2.51 g.m2) because of younger soils with low P adsorption capacity. We also found that fertilizer use significantly increased AP stocks in agricultural areas compared to native ones. Our results indicated that AP stocks strongly influenced Brazil's agricultural production, with a correlation coefficient ranging from 0.20 for coffee crops to 0.46 for soybean. The maps generated in this study are expected to contribute to the sustainable use of P in agriculture and environmental systems.

Contextually Propagated Term Weight based Approach for Author Profiling: Gender, Age and Language Variety Prediction

Author profiling creates characterization of an author based on attributes such as age, gender, language, dialect region variety, personality and so on. In recent years it has garnered significant attention for its varies applications across forensic linguistics, marketing, cybersecurity and social media analytics. Most of the research focused on stylistic, content based and term weight measures based feature representations. We observed that context semantics is not considered in the feature representation. In this propose contextually propagated term weight measures for feature representation. We implemented SVM, Random Forest and XG Boost machine learning algorithms on those feature representations. The results demonstrated that the proposed contextually propagated term weight with inverse category frequency outperformed the existing methods.

Groundwater Level Forecasting Using Machine Learning: A Case Study of the Baekje Weir in Four Major Rivers Project, South Korea

AbstractUnderstanding the impact of human‐made structures on groundwater levels is essential, with structures like dams or weirs presenting unique challenges and opportunities for study. The Baekje weir in South Korea presents an interesting case as the weir has undergone full gate opening, which is generally not the case for weirs and reservoirs, providing valuable opportunity for simulating weir removal conditions. The main objectives are investigation of groundwater level fluctuations under various weir operations, distances from the weir, and seasonal variations. The study utilizes observed data that simulates conditions with and without the weir, including scenarios of full gate opening. Multiple machine learning algorithms—Random Forest (RF), Artificial Neural Network, Support Vector Regression (SVR), Gradient Boosting, and Extreme Gradient Boosting (XGBoost)—are used to develop accurate groundwater level prediction models. The models' performance is assessed using coefficient of determination, Root mean square error (RMSE), Mean Absolute Error (MAE) indices, and visualized through Taylor diagrams. Results indicate that XGBoost outperforms other models in all three groups during both training and testing phases. Specifically, XGBoost surpasses RF by 2.09% (R2), 5.66% (RMSE), and 10.1% (MAE) in training, and outperforms SVR by 11.2% (R2), 42.0% (RMSE), and 129.2% (MAE) in testing. Additionally, the study generates groundwater level maps, providing a practical tool for managing groundwater systems and informing decision‐making in weir operations. This study not only sheds light on the dynamic relationship between weir operations and groundwater levels but also provides actionable insights for effective water management in similar hydrological settings.

Comparison and analysis of prediction accuracy between traditional machine learning algorithms and XGBoost algorithm in music emotion classification

Music emotion classification refers to determining the emotional types of music, such as happiness, sadness, anger, and passion, based on aspects like rhythm, melody, and tone. The development of artificial intelligence can be applied to music recommendation by employing machine learning algorithms to ascertain emotional attributes of music, enabling precise music suggestions for users to enhance their musical experience. This paper compares and analyzes the classification performance of traditional Random Forest machine learning algorithms and the XGBoost model on a Turkish music emotion dataset. We selected a comprehensive dataset from Kaggle, containing a vast array of music samples and their corresponding emotional labels, making it an extensive music emotion classification dataset. From the experimental results, it's evident that the traditional Random Forest machine learning model outperforms the XGBoost model in terms of accuracy, precision, and recall. The accuracy of the traditional Random Forest machine learning model stands at 80.8%, whereas the XGBoost model's accuracy is 75%. The recall rate for the traditional Random Forest machine learning model is 80.8%, while for XGBoost, it's 77.2%. The F1 score for the traditional Random Forest machine learning model is 80.5%, whereas for XGBoost, it's 75.3%. These parameters indicate that the traditional Random Forest machine learning model exhibits superior predictive performance in music emotion classification. However, the XGBoost model possesses its own advantages, such as faster learning and prediction speeds, high accuracy, and strong generalization capabilities. In summary, the traditional Random Forest machine learning model demonstrates better robustness and interpretability, effectively handling samples with noise and missing data, thus finding widespread practical application. On the other hand, the XGBoost model excels in rapid training and prediction, coupled with higher accuracy and versatility, making it advantageous in dealing with large datasets. The research outcomes of this paper hold significant importance for the study and application of music emotion classification. The experimental results presented herein offer valuable insights for researchers and practitioners, aiding them in selecting appropriate machine learning models, optimizing, and adjusting them to achieve the best classification results.

Fall Detection System based on iOS Smartphone Sensors

Activities of daily living (ADLs) are a crucial aspect of human life, especially in remote health monitoring and fall detection. As smartphones have become an integral part of our daily routines, with their ability to perform complex calculations, connect to the internet, and incorporate various sensors, researchers have been inspired to explore human activity recognition systems. This paper focuses on accelerometer and gyroscope data from iOS-based smartphones. We developed a data collection app to record fall types (e.g., Falling Right, Falling Left) and fall-like activities (e.g., Sitting Fast, Jumping). Volunteers carried smartphones naturally in their pockets during experiments, presenting a challenge of noise but enhancing user comfort. We applied different Machine Learning algorithms (Decision Trees, Random Forest, Logistic Regression, k-Nearest Neighbor, XGBoost, LightGBM, and Neural Networks) to analyze the collected dataset. In contrast to typical studies, our approach replicated real-world smartphone usage. The paper presents and analyzes promising results from the study. Furthermore, we implemented the trained model as a real-time mobile application for potential users. This research illustrates the potential of smartphones in fall detection and opens the way for user-friendly solutions in remote health monitoring.

Class imbalance: A crucial factor affecting the performance of tea plantations mapping by machine learning

Due to disparities in area among various land cover types, class imbalance has always existed in crop mapping research, posing uncertainties in extracting minority classes which occupy a smaller area. In this paper, taking tea plantations mapping in Hangzhou city as an example, we created a series of training datasets with different imbalance-ratios (IRs), compared the accuracy between the extraction models using these datasets, and analyzed the impact of class imbalance on various machine learning algorithms (Artificial Neural Network, Decision Tree, Random Forest and XGBoost), aiming to provide a feasible approach to improve the mapping accuracy of minority classes. The leave-one-out cross validation results showed that in most cases, with the increase of the IR, the model’s F2-score first increased and then decreased, and the increase of F2-scores ranged from 0.2% to 29.2%, suggesting that moderately increasing the number of other samples in the training dataset can improve the tea plantations extraction accuracy. Consistent result can also be obtained by using the whole city’s samples for modeling and random sampling validation. XGBoost performed best among the four algorithms, which yielded the optimal tea plantations map with a PA of 97%, UA of 93% and F2-score of 96% when the IR of the training dataset was 6. The UA was 19% higher than that of the model using a balanced dataset (IR=1) and was 11% higher than that of the model using pseudo-balanced datasets created by the oversampling method. The conclusions of this study offer insights for the identification of minority classes, contributing to achieving higher accuracy in remote sensing crop mapping.

Intelligent prediction model of a polymer fracture grouting effect based on a genetic algorithm-optimized back propagation neural network

Polymer grouting can effectively improve the stability of surrounding rock fractures. However, in practical construction, it is difficult to judge the degree of coupling between the slurry and the rock, and the effective grouting range after grouting. Therefore, early prediction of the effect of grouting on the surrounding rock is crucial. In this paper, a new artificial intelligence method is proposed to predict the polymer fracture grouting effect. The genetic algorithm optimized back propagation neural network (GA-BP) is employed to construct an intelligent prediction model. To acquire a substantial dataset for constructing the model, an easily assembled/disassembled test apparatus for polymer fracture grouting is designed. The maximum coupling degree of the fractures and slurry diffusion distance are chosen as the evaluation metrics for the grouting effectiveness. The influences of the fracture characteristic parameters and grouting volume on the grouting effect are investigated. Furthermore, a comprehensive analysis is conducted on the spatiotemporal diffusion characteristics and slurry-rock coupling mechanism of polymer grouting. Compared to traditional BP neural networks, and three other machine learning algorithms (decision trees, random forests and gradient boosting decision trees), the GA-BP model outperforms them in terms of R2 (coefficient of determination), MSE (mean squared error), MBE (mean bias error), MAE (mean absolute error) and RMSE (root mean squared error) in both the test and training sets. The GA algorithm significantly improves the accuracy and robustness of the prediction model. The optimized model demonstrates significant accuracy in predicting grouting results and assessing efficiency, providing a practical reference for grouting construction.

Further adoption of conservation tillage can increase maize yields in the western US Corn Belt

Conservation tillage can reduce soil erosion, increase soil health, and decrease labor and fuel input costs. Despite these benefits, potential yield impacts remain an important concern for farmers considering adoption. Previous research suggests that conservation tillage is likely to have the largest yield benefits in more arid conditions, but a lack of field-level analyses across climatic, management and soil conditions limits confidence in such predictions. Satellite imagery provides the opportunity to monitor agricultural lands at sub-field resolution across large spatial scales and wide environmental gradients. Here we investigate the maize yield impacts of conservation tillage in the semi-arid western US Corn Belt, using sub-field resolution datasets on tillage practices and crop yields derived from satellite data spanning four states (Nebraska, Kansas, South Dakota, and North Dakota) between 2008 and 2020. On these datasets, we estimate heterogenous yield outcomes for several thousand maize fields across gradients in climate, soil quality and irrigation status by using a causal forests analysis, an adaptation of the random forests machine-learning algorithm for causal inference on observational data. We find that long-term adoption of conservation tillage increased rainfed maize yields by an average of 9.9% in the region. Impacts on irrigated yields were small and not statistically significant. These results, along with an analysis of variables related to greater than average yield benefits, indicate that improved water infiltration and retention are the primary reasons for conservation tillage benefits. Despite yield benefits, many fields estimated to see increased yields under long term low till have not adopted the practice. Therefore, we identify specific counties likely to benefit most from increased levels of adoption. Our results strengthen the understanding of the impacts of conservation agriculture on crop yields and help define environments and counties most likely to benefit from conservation tillage.

Effectiveness of the method for selecting rotary-steerable systems based on the machine learning algorithm Random Forest Classifier

Relevance. Urgent need to consider and determine possible ways to use machine learning methods in drilling industry, since artificial intelligence is developing rapidly. Achieving this task will provide industrial enterprises with a huge competitive advantage and make an important contribution to the scientific community for its future research. This is emphasized by such regulations as the Decree of the President of the Russian Federation dated 10.10.2019 G. No. 490 "On the development of Artificial Intelligence in the Russian Federation" and "The National Strategy for the Development of Artificial Intelligence for the period up to 2030". Aim. To study the effectiveness of using the machine learning method Random Forest Classifier, to develop methods for selecting rotary-steerable systems, to consider the efficiency of machine learning to determine target parameters when solving the task assigned to it within the drilling industry and to determine the approximate amount of time that can be spent by the algorithm to work out a possible solution. Object. Random Forest Classifier machine learning method in the conditions of solving a problem from the drilling industry on the selection of an optimal rotary-steerable system for specifically specified conditions. Methods. The authors have performed two computational experiments using two computing and electronic machines, namely a laptop and a remote server, the prerequisite for which was the data collected and analyzed on the basis of the study of the scientific literature in the field of research. This article explores the possibility of using the machine learning method Random Forest Classifier, to optimize well construction, using the example of developing a method for selecting rotary-steerable systems. Computational experiments were performed on two computers using the Python programming language, version 3.8.10, as well as the following libraries: NumPy, Pandas, Scikit-learn. Results. The computational experiments carried out proved the ability of the considered method to solve the problems of choosing suitable drilling equipment, an example of which was rotary-steerable systems. This method is able to independently determine the dependencies necessary to perform the task and spends a small amount of time on this process. The totality of these conclusions makes it possible to unequivocally assert the expediency and necessity of developing new approaches to the use of machine learning methods in the drilling industry, as well as performing multiple scientific studies on the possibilities of using machine learning in well construction and analyzing their effectiveness, since this direction is advanced and can radically change existing ideas about the processes occurring during well drilling.

Intelligent Low-Consumption Optimization Strategies: Economic Operation of Hydropower Stations Based on Improved LSTM and Random Forest Machine Learning Algorithm

The economic operation of hydropower stations has the potential to increase water use efficiency. However, there are some challenges, such as the fixed and unchangeable flow characteristic curve of the hydraulic turbines, and the large number of variables in optimal load distribution, which limit the progress of research. In this paper, we propose a new optimal method of the economic operation of hydropower stations based on improved Long Short-Term Memory neural network (I-LSTM) and Random Forest (RF) algorithm. Firstly, in order to accurately estimate the water consumption, the LSTM model’s hyperparameters are optimized using improved particle swarm optimization, and the I-LSTM method is proposed to fit the flow characteristic curve of the hydraulic turbines. Secondly, the Random Forest machine learning algorithm is introduced to establish a load-distribution model with its powerful feature extraction and learning ability. To improve the accuracy of the load-distribution model, we use the K-means algorithm to cluster the historical data and optimize the parameters of the Random Forest model. A Hydropower Station in China is selected for a case study. It is shown that (1) the I-LSTM method fits the operating characteristics under various working conditions and actual operating characteristics of hydraulic turbines, ensuring that they are closest to the actual operating state; (2) the I-LSTM method is compared with Support Vector Machine (SVM), Extreme Learning Machine (ELM) and Long Short-Term Memory neural network (LSTM). The prediction results of SVM have a large error, but compared with ELM and LSTM, MSE is reduced by about 46% and 38% respectively. MAE is reduced by about 25% and 21%, respectively. RMSE is reduced by about 27% and 24%, respectively; (3) the RF algorithm performs better than the traditional dynamic programming algorithm in load distribution. With the passage of time and the increase in training samples, the prediction accuracy of the Random Forest model has steadily improved, which helps to achieve optimal operation of the units, reducing their average total water consumption by 1.24%. This study provides strong support for the application of intelligent low-consumption optimization strategies in hydropower fields, which can bring higher economic benefits and resource savings to renewable energy production.

Factors to improve odds of success following medial opening-wedge high tibial osteotomy: a machine learning analysis

BackgroundAlthough high tibial osteotomy (HTO) is an established treatment option for medial compartment osteoarthritis, predictive factors for HTO treatment success remain unclear. This study aimed to identify informative variables associated with HTO treatment success and to develop and internally validate machine learning algorithms to predict which patients will achieve HTO treatment success for medial compartmental osteoarthritis.MethodsThis study retrospectively reviewed patients who underwent medial opening-wedge HTO (MOWHTO) at our center between March 2010 and December 2015. The primary outcomes were a lack of conversion to total knee arthroplasty (TKA) and achievement of a minimal clinically important difference of improvement in the Knee Injury and Osteoarthritis Outcome Score (KOOS) at a minimum of five years postoperatively. Recursive feature selection was used to identify the combination of variables from an initial pool of 25 features that optimized model performance. Five machine learning algorithms (XGBoost, multilayer perception, support vector machine, elastic-net penalized logistic regression, and random forest) were trained using five-fold cross-validation three times and applied to an independent test set of patients. The performance of the model was evaluated by the area under the receiver operating characteristic curve (AUC).ResultsA total of 231 patients were included, and 200 patients (86.6%) achieved treatment success at the mean of 9 years of follow-up. A combination of seven variables optimized algorithm performance, and the following specific cutoffs increased the likelihood of MOWHTO treatment success: body mass index (BMI) ≤ 26.8 kg/m2, preoperative KOOS for pain ≤ 46.0, preoperative KOOS for quality of life ≤ 33.0, preoperative International Knee Documentation Committee score ≤ 42.0, preoperative Short-Form 36 questionnaire (SF-36) score > 42.25, three-month postoperative hip-knee-ankle angle > 1.0°, and three-month postoperative medial proximal tibial angle (MPTA) > 91.5° and ≤ 94.7°. The random forest model demonstrated the best performance (F1 score: 0.93; AUC: 0.81) and was transformed into an online application as an educational tool to demonstrate the capabilities of machine learning.ConclusionsThe random forest machine learning algorithm best predicted MOWHTO treatment success. Patients with a lower BMI, poor clinical status, slight valgus overcorrection, and postoperative MPTA < 94.7 more frequently achieved a greater likelihood of treatment success.Level of evidenceLevel III, retrospective cohort study.