Single Prediction Model Research Articles

Conv‐ELSTM: An ensemble deep learning approach for predicting short‐term wind power

AbstractAccurate and reliable forecasting of wind power is essential for the stable integration of wind energy into the electrical grid. However, the chaotic nature of wind power presents a significant challenge in utilizing data for effective short‐term forecasting, such as 60‐min predictions. This article introduces a hybrid data‐driven framework that employs an ensemble deep learning model to provide highly precise short‐term wind power predictions. The framework leverages a data‐driven approach to identify the intrinsic components of wind power data, including high‐frequency and low‐frequency components. A convolutional layer‐based feature fusion network is then established to properly extract important information from irrelevant wind energy features. Subsequently, an ensemble of long short‐term memory (LSTM) networks is developed to forecast wind power using the fused features, thereby mitigating the disadvantage of a single prediction model. The numerical experiment is carried out based on two different real‐life datasets. The results demonstrate the effectiveness of the proposed method in forecasting short‐term wind power compared to five benchmarks.

A River Water Quality Prediction Method Based on Dual Signal Decomposition and Deep Learning

Traditional single prediction models struggle to address the complexity and nonlinear changes in water quality forecasting. To address this challenge, this study proposed a coupled prediction model (RF-TVSV-SCL). The model includes Random Forest (RF) feature selection, dual signal decomposition (Time-Varying Filtered Empirical Mode Decomposition, TVF-EMD, and Sparrow Search Algorithm-Optimized Variational Mode Decomposition, SSA-VMD), and a deep learning predictive model (Sparrow Search Algorithm-Convolutional Neural Network-Long Short-Term Memory, SSA-CNN-LSTM). Firstly, the RF method was used for feature selection to extract important features relevant to water quality prediction. Then, TVF-EMD was employed for preliminary decomposition of the water quality data, followed by a secondary decomposition of complex Intrinsic Mode Function (IMF) components using SSA-VMD. Finally, the SSA-CNN-LSTM model was utilized to predict the processed data. This model was evaluated for predicting total phosphorus (TP), total nitrogen (TN), ammonia nitrogen (NH3-N), dissolved oxygen (DO), permanganate index (CODMn), conductivity (EC), and turbidity (TB), across 1, 3, 5, and 7-d forecast periods. The model performed exceptionally well in short-term predictions, particularly within the 1–3 d range. For 1-, 3-, 5-, and 7-d forecasts, R2 ranged from 0.93–0.96, 0.79–0.87, 0.63–0.72, and 0.56–0.64, respectively, significantly outperforming other comparison models. The RF-TVSV-SCL model demonstrates excellent predictive capability and generalization ability, providing robust technical support for water quality forecasting and pollution prevention.

SPLENDID incorporates continuous genetic ancestry in biobank-scale data to improve polygenic risk prediction across diverse populations.

Polygenic risk scores are widely used in disease risk stratification, but their accuracy varies across diverse populations. Recent methods large-scale leverage multi-ancestry data to improve accuracy in under-represented populations but require labelling individuals by ancestry for prediction. This poses challenges for practical use, as clinical practices are typically not based on ancestry. We propose SPLENDID, a novel penalized regression framework for diverse biobank-scale data. Our method utilizes ancestry principal component interactions to model genetic ancestry as a continuum within a single prediction model for all ancestries, eliminating the need for discrete labels. In extensive simulations and analyses of 9 traits from the All of Us Research Program (N=224,364) and UK Biobank (N=340,140), SPLENDID significantly outperformed existing methods in prediction accuracy and model sparsity. By directly incorporating continuous genetic ancestry in model training, SPLENDID stands as a valuable tool for robust risk prediction across diverse populations and fairer clinical implementation.

Landslide displacement prediction using time series InSAR with combined LSTM and TCN: application to the Xiao Andong landslide, Yunnan Province, China

AbstractResearch on landslide displacement prediction based on interferometric synthetic aperture radar (InSAR) deformation data involves two main issues. First, InSAR can provide only one-dimensional deformation data along the satellite’s line of sight (LOS), which cannot truly reflect the deformation of the landslide body in the downward direction along the slope. Second, the use of a single prediction model does not adequately account for both long-term and local changes in landslide displacement, affecting the accuracy of the predictions. To address this, in this study, Long Short-Term Memory networks (LSTM) and temporal convolutional network (TCN) models are combined to construct a method (LSTM-TCN) of landslide displacement prediction. This method can consider the long-term and localized changes in landslide displacement. The method is first based on InSAR technology to obtain surface deformation. The deformation of the landslide is subsequently computed in the downward direction along the slope to obtain the landslide displacement time series data. Next, the LSTM-TCN is used for landslide displacement prediction. Finally, the mean absolute error (MAE), root mean square error (RMSE) and coefficient of determination (R2) are used to evaluate the performance of the model. The experiment is conducted on the Xiao Andong landslide in Anshi village, Fengqing County, Lincang City, Yunnan Province, China. The LSTM-TCN model achieves an R2 of 0.75, an RMSE of 0.43 cm, and an MAE of 0.36 cm. Compared with the individual LSTM and TCN models, the LSTM-TCN model exhibits the highest prediction accuracy and the smallest prediction error, which is closer to the true result that in the other models. These results demonstrate that the combined LSTM-TCN model effectively captures the complex features and long-term trends in landslide displacement data, significantly enhancing the accuracy of predictions.

Distributed photovoltaic power forecasting based on personalized federated adversarial learning

Existing distributed photovoltaic (PV) power forecasting methods fail to address the impact of sample scarcity and heterogeneity in PV power data. Moreover, training a single prediction model proves challenging to meet the personalized forecasting needs of different PV stations in distributed environments. This paper proposes a personalized federated generative adversarial network (PFedGAN)-based DPV power forecasting method. Leveraging the federated learning (FL) framework, it achieves collaborative training of prediction models among DPV stations while preserving data privacy. y introducing generative adversarial networks (GAN) and personalized strategy optimization into the FL training process, it alleviates data scarcity issues and reduces the impact of data heterogeneity. A TimesNet-DeepAR (TNE-DeepAR) power prediction model is designed, where the TimesNet module extracts correlations between PV power data from different time periods, and the DeepAR module facilitates PV power prediction, mitigating the effects of meteorological factors' multi-periodic variations on PV power. Experimental results show that the proposed hybrid prediction model reduces the average mean absolute percentage error (MAPE) by 30–40 % compared to single models. The proposed approach reduces the MAPE by 9.79 % compared to traditional methods and by 49.62 % for PV stations with scarce data.

Traffic Flow Prediction: A Method Using Bagging-Based Ensemble Learning Model

For the development of the national economy, transportation is strategically significant. As the increasing ownership of automobiles, traffic jams are a common occurrence. Accurate prediction of traffic flow contributes to diverting traffic effectively and improving the quality of urban traffic, in turn improving the operation of the overall transportation system. The rapid development of artificial intelligence technologies, especially machine learning and deep learning, has provided effective methods for accurate prediction of traffic flow. Based on the above, in order to improve the accuracy of the prediction and to extend the application of machine learning and deep learning in the prediction of traffic flow, this study proposed a bagging-based ensemble learning model. Firstly, normalization method is used to preprocess the data. Subsequently, base prediction models including decision tree, random forest, logistic regression, convolution neural network, long short-term memory and multilayer perceptron are selected for training the prediction model, respectively. Finally, bagging-based ensemble learning method is used to integrate these base prediction models to further predict traffic flow. The results of comparison between the single base prediction models and the bagging-based ensemble learning model on the five evaluation indicators show that, for predicting the traffic flow, the bagging-based ensemble learning model outperforms the base prediction models. Meanwhile, this study explores the potential in the application of machine learning, deep learning, and especially bagging-based ensemble learning to predict traffic flow.

CNN-BiLSTM-Attention Based Stock Price Prediction and Quantitative Investment Strategy

With the rapid development of technology and the increasing maturity of financial markets, stock price prediction has become a hot topic and an important trend in the financial field. However, the stock market has complexity and volatility, in order to reduce the investment risk and ensure the maximization of benefits, selecting the optimal stock to predict the stock price and formulating the quantitative trading strategy are extremely important issues for financial academics and investors. For the research of stock price prediction and quantitative trading, firstly, the quantitative stock selection model combining XGBoost and multi-factor stock selection model is constructed, and four stocks are screened out, and then the CNN-BiLSTM-Attention model is proposed to predict the stock price trend of the selected stocks, and it is found from the prediction results of the four stocks that the prediction accuracies all reach more than 95%, which is higher than that of the single model prediction. accuracy and passed the validity test of the fitted model. Secondly, based on the quantitative investment portfolio given in the above analysis and the quantitative stock trading strategy of MACD, the prediction results are verified, and the total return of most stocks based on the strategy is higher than 40% at the highest, and the loss is controlled within 10%, which indicates that the trading strategy in this paper is effective and feasible. The study concludes that greater returns can be obtained by calculating the total returns of stock portfolios based on different portfolio approaches.

A novel high-precision and self-adaptive prediction method for ship energy consumption based on the multi-model fusion approach

The accurate prediction of energy consumption is significant for ship energy efficiency optimization. However, the existing prediction methods of ship energy consumption based on a single algorithm have limitations in adaptability and accuracy. Therefore, a novel high-precision and self-adaptive prediction method based on the multi-model fusion approach is investigated in this paper. Firstly, the data processing and analysis are carried out. Then, the Stacking-based fusion model is established and the prediction performance is analyzed. On this basis, an adaptive fusion model based on the Self-adaptive Parameter Optimization (SPO) method is established. Finally, an Intelligent Selection based Self-adaptive Hybrid (ISSH) method is proposed. The study results indicate that the proposed ISSH method can predict ship energy consumption more accurately, with the Mean Square Error (MSE) reduced by 66.7 % and the Mean Absolute Error (MAE) reduced by 12.7 % compared to the optimal single prediction model. In addition, the ISSH-based fusion model can reduce the MSE by 50.0 % and the MAE by 9.9 %, compared to the Stacking-based fusion model without parameter optimization. Moreover, the ISSH method can achieve self-adaptive prediction of ship energy consumption under diverse scenarios by adopting the intelligent selection strategy (ISS) method of basic models, which is significant to achieve dynamic optimization of ship energy efficiency.

A Risk Prediction Model for New-Onset Chronic Kidney Disease in the Elderly

Plain Language SummaryDifferent regions have different risk factors for chronic kidney disease (CKD), and using a single model for prediction can often lead to errors. This study aimed to create an easy-to-use tool to predict the risk of CKD in older adults. We analyzed data from 5,416 people aged over 65 who had regular health check-ups between January 2017 and July 2021. We found that about 20% developed CKD during a follow-up period of 2.3 years. Factors like age, being female, diabetes, high triglyceride levels, and initial kidney function were used to build the prediction model. This model was good at forecasting CKD risk and was validated to ensure its accuracy. It can help doctors quickly identify older adults who might be at risk of kidney problems and start early treatment.

Predicting axillary lymph node metastasis in breast cancer based on ultrasound radiofrequency time-series analysis.

The status of axillary lymph nodes (ALN) plays a critical role in the management of patients with breast cancer. It is an urgent demand to develop highly accurate, non-invasive methods for predicting ALN status. To evaluate the efficacy of ultrasound radiofrequency (URF) time-series parameters, in combination with clinical data, in predicting ALN metastasis in patients with breast cancer. We prospectively gathered clinicopathologic and ultrasonic data from patients diagnosed with breast cancer. Various machine-learning (ML) models were developed using all available features to determine the most efficient diagnostic model. Subsequently, distinct prediction models were created using the optimal ML model, and their diagnostic performances were evaluated and compared. The study encompassed 240 patients, of whom 88 had lymph node metastases. A leave-one-out cross-validation (LOOCV) method was used to split the entire dataset into training and testing subsets. The random forest ML model outperformed the other algorithms, with an area under the curve (AUC) of 0.92. Prediction models based on clinical, ultrasonic, URF parameters, clinical + ultrasonic, clinical + URF, and ultrasonic + URF parameters had AUCs of 0.56, 0.79, 0.78, 0.90, 0.80, and 0.84, respectively, in the testing set. The comprehensive diagnostic model (clinical + ultrasonic + URF parameters) demonstrated strong diagnostic capability, with an AUC of 0.94 in the testing set, exceeding any single prediction model. The combined model (clinical + ultrasonic + URF parameters) could be used preoperatively to predict lymph node status, offering valuable input for the design of individualized surgical approaches.

Integrating plasma exosomal miRNAs, ultrasound radiomics and tPSA for the diagnosis and prediction of early prostate cancer: a multi-center study.

This multi-center study aims to explore the roles of plasma exosomal microRNAs (miRNAs), ultrasound (US) radiomics, and total prostate-specific antigen (tPSA) levels in early prostate cancer detection. We analyzed the publicly available dataset GSE112264 to identify the differentially expressed miRNAs associated with prostate cancer. Then, PyRadiomics was used to extract image features, and least absolute shrinkage and selection operator (LASSO) was used to screen the data. Subsequently, according to strict inclusion and exclusion criteria, the internal dataset (n = 199) was used to construct a diagnostic model, and the receiver operating characteristic (ROC) curve, calibration curve, decision curve analysis (DCA), and DeLong test were used to evaluate its diagnostic performance. Finally, we used an external dataset (n = 158) for further validation. The number of features extracted by PyRadiomics was 851, and the number of features screened by LASSO was 23. We combined the hsa-miR-320c, hsa-miR-944, radiomics, and tPSA features to construct a joint model. The area under the ROC curve of the combined model was 0.935. In the internal validation, the area under the curve (AUC) of the training set was 0.943, and the AUC of the test set was 0.946. The AUC of the external data set was 0.910. The calibration curve and decision curve were consistent with the performance of the combined model. There was a significant difference in the prediction ability between the combined prediction model and the single index prediction model, indicating the high credibility and accuracy of the combined model in predicting PCa. The combined prediction model, consisting of plasma exosomal miRNAs (hsa-miR-320c and hsa-miR-944), US radiomics, and clinical tPSA, can be utilized for the early diagnosis of prostate cancer.

Predicting oral and esophageal cancers by one model in a Chinese prospective cohort study

ObjectiveOral and esophageal cancers are both upper gastrointestinal cancers that share a number of risk factors. However, most previous risk prediction models only focused on one of these two types of cancer. There is no single model that could predict both cancers simultaneously. Our objective was to develop a model specifically tailored for oral and esophageal cancers. MethodsFrom 1996 to 2007, a total of 431,460 subjects aged 20 and older without a history of cancer at baseline were included and were monitored for an average duration of 7.3 years in Taiwan, China. A total of 704 cases of oral and esophageal cancers were detected. We utilized both univariate and multivariate COX regression for screening predictors and constructing the model. We evaluated the goodness of fit of the model based on discriminatory accuracy, Harrell's C-index, and calibration. ResultsFinally, we developed a Cox regression model using the twelve most significant variables: age, gender, alcohol consumption, betel chewing, smoking status, history of oral ulceration, educational level, marital status, oropharynx status, family history of nasopharyngeal carcinoma, volume ratio of blood cell, and gamma-glutamyl transferase. The AUC (area under the curve) for the complete model was 0.82. Additionally, the C-index was 0.807 (with a 95 % confidence interval ranging from 0.789 to 0.824) and internal calibration results demonstrated that the model performed well. ConclusionsThis study identified the twelve most significant common risk factors for oral and esophageal cancers and developed a single prediction model that performs well for both types of cancer.

High-speed railway express delivery volume forecast based on data-driven ensemble forecast approaches: The China case

Current researches on logistics delivery volume forecast mainly focus on traditional transportation modes such as road, railway, and aviation, with little research on predicting high-speed railway express delivery (HSRED) volume. In this paper, a data-driven ensemble forecast approach is proposed to forecast the HSRED. Firstly, we use four kinds of single methods to predicate the HSRED, include Auto-Regressive Moving Average Model (ARMA), Linear Regression (LR), Random Forest (RF) and Singular Spectrum Analysis (SSA). Secondly, based on the predication results of single methods, we apply two ensemble methods including Arithmetic Mean (AM) and Allocation Integration Operator (AIO) to determine 22 combinations of ensemble forecast approaches. The Mean Absolute Deviation (MAD), Mean Absolute Percentage Error (MAPE), Root Mean Squared Error (RMSE) and Correlation Coefficient (CC) are used to evaluate the performance of each single and ensemble approach. Next, a real-world case study is conducted by using the HSRED in China as the background, the delivery demand data from Jan. 2014 to Mar. 2023 is used as the input. The results show that: The prediction results based on SSA, ARMA, and RF are close to the actual data. LR has the worst predication performance. The overall performance indicators of the ensemble models are better than those of the single prediction models. The ensemble models by using AIO are generally better than those by using AM. SSA-ARMA ensemble model constructed by AIO shows the best predictive performance. The trend of HSRED changes from 2025 to 2050 is gentle, and there will be no explosive growth, and it will gradually decrease and eventually stabilize. Finally, we distribute the predicted results to each province based on the gravity model and conduct visual analysis. The results show that: Provinces with high demand for HSRED include Shandong, Zhejiang, Jiangsu, Shanghai, and Beijing, mostly concentrated in the eastern and southeastern regions, we can consider strengthening the development and increasing investment of HSRED in these cities. Provinces with less demand, efforts can be made to strengthen publicity, accelerate the formation process of the high-speed railway network, and increase support for the HSRED industry.

Research on Agricultural Product Price Prediction Based on Improved PSO-GA

The accurate prediction of scallion prices can not only optimize supply chain management and help related practitioners and consumers to make more reasonable purchasing decisions, but also provide guidance for farmers’ planting choices, thus enhancing market efficiency and promoting the sustainable development of the whole industry. This study adopts the idea of decomposition–denoising–aggregation, using three decomposition and denoising techniques combined with three single prediction models to form a base model. Various base models are divided into different combinations based on whether the computational structure is the same or not, and the optimal weights of the combinations are determined by using the improved particle swarm optimization–genetic algorithm (PSO-GA) optimization algorithm in different combinations. The experimental results show that the scallion price in Shandong Province from 2014 to 2023 shows an overall upward trend, and there is a cyclical and seasonal fluctuation pattern of “high in winter and low in summer”; the semi-heterogeneous-PSO-GA model reduces the MAPE by 49.03% and improves the directional accuracy by 41.52%, compared to the optimal single prediction model, ARIMA. In summary, the combined model has the most accurate prediction and strong robustness, which can provide ideas and references for the difficult problem of determining the optimal weights of the combined model in the field of predicting the prices of agricultural products.

Computed tomography-based radiomics and clinical-genetic features for brain metastasis prediction in patients with stage III/IV epidermal growth factor receptor-mutant non-small-cell lung cancer.

To evaluate the value of computed tomography (CT)-based radiomics combined with clinical-genetic features in predicting brain metastasis in patients with stage III/IV epidermal growth factor receptor (EGFR)-mutant non-small-cell lung cancer (NSCLC). The study included 147 eligible patients treated at our institution between January 2018 and May 2021. Patients were randomly divided into two cohorts for model training (n = 102) and validation (n = 45). Radiomics features were extracted from the chest CT images before treatment, and a radiomics signature was constructed using the Least Absolute Shrinkage and Selection Operator regression. Kaplan-Meier survival analysis was used to describe the differences in brain metastasis-free survival (BM-FS) risk. A clinical-genetic model was developed using Cox regression analysis. Radiomics, genetic, and combined prediction models were constructed, and their predictive performances were evaluated by the concordance index (C-index). Patients with a low radiomics score had significantly longer BM-FS than those with a high radiomics score in both the training (p < 0.0001) and the validation (p = 0.0016) cohorts. The C-indices of the nomogram, which combined the radiomics signature and N stage, overall stage, third-generation tyrosine kinase inhibitor treatment, and EGFR mutation status, were 0.886 (95% confidence interval [CI] 0.823-0.949) and 0.811 (95% CI 0.719-0.903) in the training and validation cohorts, respectively. The combined model achieved a higher discrimination and clinical utility than the single prediction models. The combined radiomics-genetic model could be used to predict BM-FS in stage III/IV NSCLC patients with EGFR mutations.

PBPK model-based gender-specific risk assessment of N-nitrosodimethylamine (NDMA) using human biomonitoring data.

Despite several screening levels for NDMA reported in water, soil, air, and drugs, the human risk assessment using biomonitoring concentrations has not been performed. In this study, gender-specific exposure guidance values were determined in humans, then biomonitoring measurements in healthy Korean subjects (32 men and 40 women) were compared to the exposure guidance values to evaluate the current exposure level to NDMA. For the human risk assessment of NDMA, the gender-specific physiologically based pharmacokinetic (PBPK) model was developed in humans using proper physiological parameters, partition coefficients, and biochemical parameters. Using the PBPK model, a Monte Carlo simulation was performed to describe the magnitudes of inter-individual variability and uncertainty on the single model predictions. The PBPK modeling and Monte Carlo simulation allowed the estimation of the relationship between external dose and blood concentration for the risk assessment. The procedure for the human risk assessment was summarized as follows: (1) estimating a steady-state blood concentration (Cavg) corresponding to the daily no observed adverse effect level (NOAEL) administration in rats; (2) applying uncertainty factors (UFs) for deriving the human Cavg; (3) determining the exposure guidance values as screening criteria; (4) interpreting the human biomonitoring measurements by forward and reverse dosimetry approaches. Using the biomonitoring concentrations, current daily exposures to NDMA were estimated to be 3.95μg/day/kg for men and 10.60μg/day/kg for women, respectively. The result of the study could be used as a basis for implementing further risk management and regulatory decision-making for NDMA.

Fusion prediction strategy-based dynamic multi-objective sparrow search algorithm

Solving dynamic multi-objective optimization problems with time-varying Pareto front (PF) or Pareto set (PS) is a challenging task. Such problems require algorithms to react to environmental changes and efficiently track optimal solutions. For this purpose, a dynamic multi-objective sparrow search algorithm (SSA) with fusion prediction strategy, based on difference model and kernel extreme learning machine (DMOSSA-FPS), is proposed. Given the diversity of change characteristics, a single prediction model is insufficient. Therefore, based on the historical information of the population, a difference model and a kernel extreme learning machine are integrated for PS prediction. The former is used to predict the solutions of some individuals under approximate linear changes and the latter is employed for nonlinear predictions. In a new environment, the combined predictions increase the diversity of the initial population. Additionally, a new static optimizer is proposed, which combines decomposition- and dominance-based approaches to constitute a new individual screening mechanism. Then the optimization mode of SSA is introduced to enhance both algorithmic diversity and convergence rate. The experimental results on the DF test suite demonstrate that, compared with several other advanced algorithms, DMOSSA-FPS exhibits stronger convergence and robustness.

A novel combination prediction model of ultra-short-term wind speed

The accurate prediction of ultra-short-term wind speed has important theoretical significance and practical application value. In this paper, a combination prediction model of ultra-short-term wind speed based on variational mode decomposition is proposed. Firstly, the variational mode decomposition algorithm is introduced to decompose the ultra-short-term wind speed and obtain the components of different frequency. The approximate entropy algorithm is used to calculate the complexity of each component. According to the calculation results, echo state network is selected to predict high complexity components, support vector machine is used to predict medium complexity components, and autoregressive integrated moving average model is used to predict low complexity components. Then, the predicted values of each component are added to get the final prediction result. Gray wolf algorithm is used to optimize the model parameters of support vector machine and echo state network. In addition, the approximate entropy calculation results show that compared with the original ultra-short-term wind speed, the complexity of the components obtained by the variational mode decomposition algorithm is reduced, which makes the modeling of the ultra-short-term wind speed system simpler. Finally, the validity of the model is verified by taking the ultra-short-term wind speed data of 5 minutes and 10 minutes sampling period actually collected as the research object. The results show that the prediction model proposed in this paper has better performance than other single or combination prediction models.

A forecasting model with hybrid bidirectional long short-term memory for mooring line responses of semi-submersible offshore platforms

Semi-submersible offshore platforms are often produced in the deep sea far away from land, which results in high operating costs and difficult safety maintenance due to the harsh and complex working conditions. Adopting a neural network model for online accurate prediction of mooring tension of semi-submersible offshore platforms can adjust the ballast according to the advance of mooring tension change to keep the platform in a relatively smooth movement amplitude and improve production efficiency. In addition, it can also prevent catastrophic accidents caused by anchor chain breakage due to excessive mooring tension. However, the measured mooring tension of semi-submersible offshore platforms is characterized by complex nonlinear non-stationarity, which leads to insufficient prediction accuracy of a single prediction model. This paper develops a novel online prediction model of mooring tension for semi-submersible offshore platforms based on empirical modal decomposition (EMD), convolutional neural network (CNN), and bidirectional long and short-term memory neural network (BiLSTM). The proposed model is validated using the measured data of the first semi-submersible offshore platform in the South China Sea during operation. The experimental results show that the proposed model outperforms other comparative models currently mainstream for mooring tension prediction. In addition, the proposed method can also be used to predict complex nonlinear data in the field of ship and ocean engineering.

LSTM-XGBoost Based RR Intervals Time Series Prediction Method in Hypertensive Patients

The prediction of RR intervals in hypertensive patients can help clinicians to analyze and warn patients' heart condition. Using 8 patients' data as samples, the RR intervals of patients were predicted by long short-term memory network (LSTM) and gradient lift tree (XGBoost), and the prediction results of the two models were combined by the inverse variance method to overcome the disadvantage of single model prediction. Compared with the single model, the proposed combined model had a different degree of improvement in the prediction of RR intervals in 8 patients. LSTM-XGBoost model provides a method for predicting RR intervals in hypertensive patients, which has potential clinical feasibility.