Linear Learning Model Research Articles

Reconstructing GRACE-derived terrestrial water storage anomalies with in-situ groundwater level measurements and meteorological forcing data

Study regionNorth China Plain (NCP), China, a semi-arid region with intense groundwater withdrawals. Study focusThis paper developed a framework using meteorological data, model-simulated terrestrial water storage anomalies (TWSA), and additional in-situ (groundwater level, GL) data to improve the unsatisfactory GRACE-TWSA reconstruction in arid and semi-arid regions due to the intense anthropogenic influence on groundwater. The inconsistency between point-scale data (GL) and grid-scale data (GRACE-TWSA and predictors other than GL) is handled by feature extraction techniques. Moreover, to deal with temporal non-stationarity, the time series are separated into trend and detrended components, the patterns of which are further learned by linear and nonlinear machine learning models, respectively.New hydrological insights for the region: Multi-site GL observations in NCP can not only serve as validation data but also as predictors providing invaluable information on human effects for the reconstructed TWSA improvement (from 6.51 to 3.86 cm for Root Mean Square Error and from 0.56 to 0.82 for Nash-Sutcliffe Efficiency). Our results show that multi-site GL data in NCP are highly inter-correlated and can be represented by several principal components, demonstrating the strong hydraulic connectivity in NCP. We also find a significant one-month lag and linear relationship between the trends of GRACE-TWSA and GL changes in NCP. These deeper understandings of hydrologic processes have implications for enhancing the GRACE-TWSA estimations in other similar regions.

A machine learning approach to predict cellular mechanical stresses in response to chemical perturbation

Mechanical stresses generated at the cell-cell level and cell-substrate level have been suggested to be important in a host of physiological and pathological processes. However, the influence various chemical compounds have on the mechanical stresses mentioned above is poorly understood, hindering the discovery of novel therapeutics, and representing a barrier in the field. To overcome this barrier, we implemented two approaches: 1) monolayer boundary predictor and 2) discretized window predictor utilizing either stepwise linear regression or quadratic support vector machine machine learning model to predict the dose-dependent response of tractions and intercellular stresses to chemical perturbation. We used experimental traction and intercellular stress data gathered from samples subject to 0.2 or 2 μg/mL drug concentrations along with cell morphological properties extracted from the bright-field images as predictors to train our model. To demonstrate the predictive capability of our machine learning models, we predicted tractions and intercellular stresses in response to 0 and 1 μg/mL drug concentrations which were not utilized in the training sets. Results revealed the discretized window predictor trained just with four samples (292 images) to best predict both intercellular stresses and tractions using the quadratic support vector machine and stepwise linear regression models, respectively, for the unseen sample images.

Integrating machine learning and environmental variables to constrain uncertainty in crop yield change projections under climate change

Robust crop yield projections under future climates are fundamental prerequisites for reliable policy formation. Both process-based crop models and statistical models are commonly used for this purpose. Process-based models tend to simplify processes, minimize the effects of extreme events, and ignore biotic pressures, while statistical models cannot deterministically capture intricate biological and physiological processes underpinning crop growth. We attempted to integrate and overcome shortcomings in both modelling frameworks by integrating the dynamic linear model (DLM) and random forest machine learning model (RF) with nine global gridded crop models (GGCM), respectively, in order to improve projections and reduce uncertainties of maize (Zea mays L.) and soybean (Glycine max [L.] Merrill) yield projections. Our results demonstrated substantial improvements in model performance accuracy by using RF in concert with GGCM across China’s maize and soybean belt. This improvement surpasses that achieved using DLM. For maize, the GGCM+RF models increased the r values from 0.15 to 0.61–0.64–0.77 and decreased nRMSE from approximately 0.20 to 0.50–0.13–0.17 compared with using GGCM alone. For soybean, the models increased r from 0.37 to 0.70–0.54–0.70 and decreased nRMSE from 0.17 to 0.35–0.17–0.20 compared with using GGCM alone. The main factors influencing maize yield changes included chilling days (CD), crop pests and diseases (CPDs), and drought, while for soybean the primary influencing factors included CPD, tropical days (based on exceeding a maximum temperature), and drought. Our approach decreased uncertainties by 33–78% for maize and by 56–68% for soybean. The main source of uncertainty for GGCM was the crop model. For GGCM+RF, the main source of uncertainty for the 2040–2069 period was the global climate model, while the main source of uncertainty for the 2070–2099 period was the climate scenario. Our results provide a novel, robust, and pragmatic framework to constrain uncertainties in order to accurately assess the impact of future climate change on crop yields. These results could be used to interpret future ensemble studies by accounting for uncertainty in crop and climate models, as well as to assess future emissions scenarios.

Assessing and comparing fixed-target forecasts of Arctic sea ice: Glide charts for feature-engineered linear regression and machine learning models

We use “glide charts” (plots of sequences of root mean squared forecast errors as the target date is approached) to evaluate and compare fixed-target forecasts of Arctic sea ice. We first use them to evaluate the simple feature-engineered linear regression (FELR) forecasts of Diebold and Göbel (2022), and to compare FELR forecasts to naive pure-trend benchmark forecasts. Then we introduce a much more sophisticated feature-engineered machine learning (FEML) model, and we use glide charts to evaluate FEML forecasts and compare them to a FELR benchmark. Our substantive results include the frequent appearance of predictability thresholds, which differ across months, meaning that accuracy initially fails to improve as the target date is approached but then increases progressively once a threshold lead time is crossed. Also, we find that FEML can improve appreciably over FELR when forecasting “turning point” months in the annual cycle at horizons of one to three months ahead.

Metabolomic differentiation of tumor core versus edge in glioma.

Gliomas exhibit high intratumor and interpatient heterogeneity. Recently, it has been shown that the microenvironment and phenotype differ significantly between the glioma core (inner) and edge (infiltrating) regions. This proof-of-concept study differentiates metabolic signatures associated with these regions, with the potential for prognosis and targeted therapy that could improve surgical outcomes. Paired glioma core and infiltrating edge samples were obtained from 27 patients after craniotomy. Liquid-liquid metabolite extraction was performed on the samples and metabolomic data were obtained via 2D liquid chromatography-mass spectrometry/mass spectrometry. To gauge the potential of metabolomics to identify clinically relevant predictors of survival from tumor core versus edge tissues, a boosted generalized linear machine learning model was used to predict metabolomic profiles associated with O6-methylguanine DNA methyltransferase (MGMT) promoter methylation. A panel of 66 (of 168) metabolites was found to significantly differ between glioma core and edge regions (p ≤ 0.05). Top metabolites with significantly different relative abundances included DL-alanine, creatine, cystathionine, nicotinamide, and D-pantothenic acid. Significant metabolic pathways identified by quantitative enrichment analysis included glycerophospholipid metabolism; butanoate metabolism; cysteine and methionine metabolism; glycine, serine, alanine, and threonine metabolism; purine metabolism; nicotinate and nicotinamide metabolism; and pantothenate and coenzyme A biosynthesis. The machine learning model using 4 key metabolites each within core and edge tissue specimens predicted MGMT promoter methylation status, with AUROCEdge = 0.960 and AUROCCore = 0.941. Top metabolites associated with MGMT status in the core samples included hydroxyhexanoycarnitine, spermine, succinic anhydride, and pantothenic acid, and in the edge samples metabolites included 5-cytidine monophosphate, pantothenic acid, itaconic acid, and uridine. Key metabolic differences are identified between core and edge tissue in glioma and, furthermore, demonstrate the potential for machine learning to provide insight into potential prognostic and therapeutic targets.

Using machine learning to realize genetic site screening and genomic prediction of productive traits in pigs.

Genomic prediction, which is based on solving linear mixed-model (LMM) equations, is the most popular method for predicting breeding values or phenotypic performance for economic traits in livestock. With the need to further improve the performance of genomic prediction, nonlinear methods have been considered as an alternative and promising approach. The excellent ability to predict phenotypes in animal husbandry has been demonstrated by machine learning (ML) approaches, which have been rapidly developed. To investigate the feasibility and reliability of implementing genomic prediction using nonlinear models, the performances of genomic predictions for pig productive traits using the linear genomic selection model and nonlinear machine learning models were compared. Then, to reduce the high-dimensional features of genome sequence data, different machine learning algorithms, including the random forest (RF), support vector machine (SVM), extreme gradient boosting (XGBoost) and convolutional neural network (CNN) algorithms, were used to perform genomic feature selection as well as genomic prediction on reduced feature genome data. All of the analyses were processed on two real pig datasets: the published PIC pig dataset and a dataset comprising data from a national pig nucleus herd in Chifeng, North China. Overall, the accuracies of predicted phenotypic performance for traits T1, T2, T3 and T5 in the PIC dataset and average daily gain (ADG) in the Chifeng dataset were higher using the ML methods than the LMM method, while those for trait T4 in the PIC dataset and total number of piglets born (TNB) in the Chifeng dataset were slightly lower using the ML methods than the LMM method. Among all the different ML algorithms, SVM was the most appropriate for genomic prediction. For the genomic feature selection experiment, the most stable and most accurate results across different algorithms were achieved using XGBoost in combination with the SVM algorithm. Through feature selection, the number of genomic markers can be reduced to 1 in 20, while the predictive performance on some traits can even be improved compared to using the full genome data. Finally, we developed a new tool that can be used to execute combined XGBoost and SVM algorithms to realize genomic feature selection and phenotypic prediction.

Comparison of linear and non-linear machine learning models for time-dependent readmission or mortality prediction among hospitalized heart failure patients

Although many models are available to predict prognosis of heart failure patients, most tools combining survival analysis are based on proportional hazard model. Non-linear machine learning algorithms would overcome the limitation of the time-independent hazard ratio assumption and provide more information in readmission or mortality prediction among heart failure patients.The present study collected the clinical information of 1796 hospitalized heart failure patients surviving during hospitalization in a Chinese clinical center from December 2016 to June 2019. A traditional multivariate Cox regression model and three machine learning survival models were developed in derivation cohort. Uno's concordance index and integrated Brier score in validation cohort were calculated to evaluate the discrimination and calibration of different models. Time-dependent AUC and Brier score curves were plotted to assess the performance of models at different time phases.

Prediction of postoperative lung function in lung cancer patients using machine learning models.

: Surgical resection is the standard treatment for early-stage lung cancer. Since postoperative lung function is related to mortality, predicted postoperative lung function is used to determine the treatment modality. The aim of this study was to evaluate the predictive performance of linear regression and machine learning models. We extracted data from the Clinical Data Warehouse (CDW) and developed 3 sets: set Ⅰ, the linear regression model; set Ⅱ, machine learning models omitting the missing data: and set Ⅲ, machine learning models imputing the missing data. Six machine learning models, the least absolute shrinkage and selection operator (LASSO), ridge regression, ElasticNet, Random Forest, eXtreme gradient boosting (XGBoost), and the light gradient boosting machine (LightGBM) were implemented. The forced expiratory volume in one second (FEV1) measured 6 months after surgery was defined as the outcome. Five-fold cross-validation was performed for hyperparameter tuning of the machine learning models. The dataset was split into training and test datasets at a 70:30 ratio. Implementation was done after dataset splitting in set Ⅲ. Predictive performance was evaluated by R2 and mean squared error (MSE) in the 3 sets. A total of 1,487 patients were included in sets Ⅰ and Ⅲ and 896 patients were included in set Ⅱ. In set Ⅰ, the R2 value was 0.27 and in set Ⅱ, LightGBM was the best model with the highest R2 value of 0.5 and the lowest MSE of 154.95. In set Ⅲ, LightGBM was the best model with the highest R2 value of 0.56 and the lowest MSE of 174.07. The LightGBM model showed the best performance in predicting postoperative lung function.

IPMN-LEARN: A linear support vector machine learning model for predicting low-grade intraductal papillary mucinous neoplasms.

We aimed to build a machine learning tool to help predict low-grade intraductal papillary mucinous neoplasms (IPMNs) in order to avoid unnecessary surgical resection. IPMNs are precursors to pancreatic cancer. Surgical resection remains the only recognized treatment for IPMNs yet carries some risks of morbidity and potential mortality. Existing clinical guidelines are imperfect in distinguishing low-risk cysts from high-risk cysts that warrant resection. We built a linear support vector machine (SVM) learning model using a prospectively maintained surgical database of patients with resected IPMNs. Input variables included 18 demographic, clinical, and imaging characteristics. The outcome variable was the presence of low-grade or high-grade IPMN based on post-operative pathology results. Data were divided into a training/validation set and a testing set at a ratio of 4:1. Receiver operating characteristics analysis was used to assess classification performance. A total of 575 patients with resected IPMNs were identified. Of them, 53.4% had low-grade disease on final pathology. After classifier training and testing, a linear SVM-based model (IPMN-LEARN) was applied on the validation set. It achieved an accuracy of 77.4%, with a positive predictive value of 83%, a specificity of 72%, and a sensitivity of 83% in predicting low-grade disease in patients with IPMN. The model predicted low-grade lesions with an area under the curve of 0.82. A linear SVM learning model can identify low-grade IPMNs with good sensitivity and specificity. It may be used as a complement to existing guidelines to identify patients who could avoid unnecessary surgical resection.

Subseasonal Prediction of Central European Summer Heatwaves with Linear and Random Forest Machine Learning Models

Abstract Heatwaves are extreme near-surface temperature events that can have substantial impacts on ecosystems and society. Early warning systems help to reduce these impacts by helping communities prepare for hazardous climate-related events. However, state-of-the-art prediction systems can often not make accurate forecasts of heatwaves more than two weeks in advance, which are required for advance warnings. We therefore investigate the potential of statistical and machine learning methods to understand and predict central European summer heatwaves on time scales of several weeks. As a first step, we identify the most important regional atmospheric and surface predictors based on previous studies and supported by a correlation analysis: 2-m air temperature, 500-hPa geopotential, precipitation, and soil moisture in central Europe, as well as Mediterranean and North Atlantic sea surface temperatures, and the North Atlantic jet stream. Based on these predictors, we apply machine learning methods to forecast two targets: summer temperature anomalies and the probability of heatwaves for 1–6 weeks lead time at weekly resolution. For each of these two target variables, we use both a linear and a random forest model. The performance of these statistical models decays with lead time, as expected, but outperforms persistence and climatology at all lead times. For lead times longer than two weeks, our machine learning models compete with the ensemble mean of the European Centre for Medium-Range Weather Forecast’s hindcast system. We thus show that machine learning can help improve subseasonal forecasts of summer temperature anomalies and heatwaves. Significance Statement Heatwaves (prolonged extremely warm temperatures) cause thousands of fatalities worldwide each year. These damaging events are becoming even more severe with climate change. This study aims to improve advance predictions of summer heatwaves in central Europe by using statistical and machine learning methods. Machine learning models are shown to compete with conventional physics-based models for forecasting heatwaves more than two weeks in advance. These early warnings can be used to activate effective and timely response plans targeting vulnerable communities and regions, thereby reducing the damage caused by heatwaves.

Analysis of Various Models for House Price Prediction based on Machine Learning

The real estate prices prediction is a relatively popular data analysis problem. With the gradual increase in the competitiveness of the housing trading platform in recent years, and the basic model of house prices predicted is different from the factual transaction price. Therefore, it is necessary to design more accurate house prices prediction models. This article will adopt different linear regression models and machine learning models. By training these models of training Set, through the accuracy algorithm of Kaggle, the accuracy of each model is obtained. According to the analysis, this article finally discovered that the accuracy of the new model is improved than the accuracy of the basic model (i.e., new model predicts house price more precision), but it can be improved in the case of less data in the real environment. The reason may be related to economic and public opinion effects. Overall, these results shed light on guiding further exploration of house price.

Integrating planar and vertical environmental features for modelling land surface temperature based on street view images and land cover data

With ongoing urbanization, high temperatures occur more frequently in urban environments. The urban heat island (UHI) effect has serious impacts on people's living environment and health, and has become an important concern for climate changes. Existing studies on environmental effects on UHI mainly focus on aspects related to urban horizontal expansion, but investigation into the effects of vertical urban growth (e.g., tall buildings) is limited. This study proposes a comprehensive set of features to characterize the planar and vertical aspects of urban environments by using panoramic Street View images (SVIs) and land cover data, and employs multiple linear regression and machine learning models to model land surface temperature (LST). Furthermore, an explainable AI method is applied to quantify the warming/cooling effects of each individual feature on LST.The results show that the proposed vertical/planar features contribute to the estimation of LST (R2 = 0.693). Adding the vertical features (extracted from SVIs) can improve LST prediction by 9.2% compared to using only planar features. Of all the features, built-up features contribute the most to LST variation, and large trees show strong cooling effects. These results serve as a basis for sustainable urban planning and help mitigate the UHI effect in cities.

Machine learning and the cross-section of emerging market stock returns

This paper compares various machine learning models to predict the cross-section of emerging market stock returns. We document that allowing for non-linearities and interactions leads to economically and statistically superior out-of-sample returns compared to traditional linear models. Although we find that both linear and machine learning models show higher predictability for stocks associated with higher limits to arbitrage, we also show that this effect is less pronounced for non-linear models. Furthermore, significant net returns can be achieved when accounting for transaction costs, short-selling constraints, and limiting our investment universe to big stocks only.

Predicting of Credit Default by SVM and Decision Tree Model Based on Credit Card Data

With the global financial crisis and increased credit risk, default forecasting is playing an increasingly important role in every sector of the economy. Currently, there are linear models and machine learning models for predicting credit defaults. In recent years, big data risk control models are superior to traditional bank models in predicting default rates, and can also conduct business quickly and on a large scale. This paper compares the SVM and the decision tree model in the machine learning model based on the credit card loan data set, and finally evaluates the prediction effect between the two models. According to the study, the decision tree model outperforms the SVM in terms of prediction accuracy. The use of big data to conduct machine learning to predict credit conditions enables financial institutions to serve small, medium and micro enterprises that were difficult to cover by traditional finance on a large scale in the past. It is a world-class innovation in finance.

Modeling Carbon Emissions of Post-Selective Logging in the Production Forests of Ulu Jelai, Pahang, Malaysia

Harvested timber and constructed infrastructure over the logging area leave massive damage that contributes to the emission of anthropogenic gases into the atmosphere. Carbon emissions from tropical deforestation and forest degradation are the second largest source of anthropogenic emissions of greenhouse gases. Even though the emissions vary from region to region, a significant amount of carbon emissions comes mostly from timber harvesting, which is tightly linked to the selective logging intensity. This study intended to utilize a remote sensing approach to quantify carbon emissions from selective logging activities in Ulu Jelai Forest Reserve, Pahang, Malaysia. To quantify the emissions, the relevant variables from the logging’s impact were identified as a predictor in the model development and were listed as stump height, stump diameter, cross-sectional area, timber volume, logging gaps, road, skid trails, and incidental damage resulting from the logging process. The predictive performance of linear regression and machine learning models, namely support vector machine (SVM), random forest, and K-nearest neighbor, were examined to assess the carbon emission from this degraded forest. To test the different methods, a combination of ground inventory plots, unmanned aerial vehicles (UAV), and satellite imagery were analyzed, and the performance in terms of root mean square error (RMSE), bias, and coefficient of correlation (R2) were calculated. Among the four models tested, the machine learning model SVM provided the best accuracy with an RMSE of 21.10% and a bias of 0.23% with an adjusted R2 of 0.80. Meanwhile, the linear model performed second with an RMSE of 22.14%, a bias of 0.72%, and an adjusted R2 of 0.75. This study demonstrates the efficacy of remotely sensed data to facilitate the conventional methods of quantifying carbon emissions from selective logging and promoting advanced assessments that are more effective, especially in massive logging areas and various forest conditions. Findings from this research will be useful in assisting the relevant authorities in optimizing logging practices to sustain forest carbon sequestration for climate change mitigation.

Deep learning regression for quantitative LIBS analysis

One of the most promising innovation strategies for sorting and recycling post-consumer aluminium scrap is using quantitative Laser-Induced Breakdown Spectroscopy (LIBS) analysis. However, existing methods to estimate alloying element concentrations based on LIBS spectra, such as linear univariate regression and Machine Learning models, are still too limited in their performance to achieve the accuracy demanded by the industry. Therefore, this study presents novel Deep Learning approaches and compares their performance to those of traditional univariate regression and Machine Learning methods in terms of RMSE, MAE, and R2 value. For this evaluation, two sample sets of aluminium pieces are used: one containing 27 certified aluminium reference samples and the second containing 733 post-consumer scrap pieces for which the ground truth concentrations are determined by X-Ray Fluorescence (XRF).Adopting multiple loss functions, one for each element, has proven its significant value for the regression performance. It improves the results for all performance metrics in the Scrap Sample set, and the same is true for the Reference Sample set, except for the coefficient of determination of Fe, Mn and Mg. In addition, the proposed methodology considers the learning prioritisation problem to prevent that learning the concentration of the base element is prioritised over the alloying elements. Although the effect of excluding the base alloy aluminium from the learning is small and not always positive for the performance, demonstrating this effect is also considered valuable. Since the average RMSE on the prediction is just 0.02 wt% for Al and Si, and not more than 0.01 wt% for Fe, Cu, Mn, Mg, and Zn, the best-performing Deep Learning model shows promise for the future of LIBS in metal sorting applications.

Prediction of linear model on stunting prevalence with machine learning approach

An increase in the number of residents should be anticipated including in the health sector, especially the problem of stunting. Stunting in children disrupts height and lack of absorption of nutrients. Information and data drive change in many areas such as health, entertainment, economics, business, and other strategic areas. The stages carried out in this study are initiating, developing linear models, and making prediction results on linear machine learning models. The results of testing with the scikit-learn linear model with a minimum variable of 19 get the best test results, namely the polynomial regression with pipeline model with mean absolute percentage error (MAPE) 0.02, root mean square error (RMSE) 3.32, and coefficient of determination (R2) 1,00. Testing with the scikit-learn linear model with a maximum variable of 48 gets the best test results, namely the polynomial regression with pipeline model with MAPE 0.00, RMSE 3.79 and R2 1.00. Testing with the scikit-learn linear model with an average variable of 32 gets the best test results, namely the polynomial regression model with MAPE 0.01, RMSE 3.32, and R2 1.00. The results of testing with the scikit-learn linear model with the minimum, maximum, and average variables get the best test results, namely the polynomial regression with pipeline model.

Diagnostic Screening of Bovine Mastitis Using MALDI-TOF MS Direct-Spotting of Milk and Machine Learning

Novel strategies for diagnostic screening of animal and herd health are crucial to contain disease outbreaks, maintain animal health, and maximize production efficiency. Mastitis is an inflammation of the mammary gland in dairy cows, often resulting from infection from a microorganism. Mastitis outbreaks result in loss of production, degradation of milk quality, and the need to isolate and treat affected animals. In this work, we evaluate MALDI-TOF mass spectrometry as a diagnostic for the culture-less screening of mastitis state from raw milk samples collected from regional dairies. Since sample preparation requires only minutes per sample using microvolumes of reagents and no cell culture, the technique is promising for rapid sample turnaround and low-cost diagnosis. Machine learning algorithms have been used to detect patterns embedded within MALDI-TOF spectra using a training set of 226 raw milk samples. A separate scoring set of 100 raw milk samples has been used to assess the specificity (spc) and sensitivity (sens) of the approach. Of machine learning models tested, the gradient-boosted tree model gave global optimal results, with the Youden index of J = 0.7, sens = 0.89, and spc = 0.81 achieved for the given set of conditions. Random forest models also performed well, achieving J > 0.63, with sens = 0.83 and spc = 0.81. Naïve Bayes, generalized linear, fast large-margin, and deep learning models failed to produce diagnostic results that were as favorable. We conclude that MALDI-TOF MS combined with machine learning is an alternative diagnostic tool for detection of high somatic cell count (SCC) and subclinical mastitis in dairy herds.

Features from the photoplethysmogram and the electrocardiogram for estimating changes in blood pressure

There is a growing emphasis being placed on the potential for cuffless blood pressure (BP) estimation through modelling of morphological features from the photoplethysmogram (PPG) and electrocardiogram (ECG). However, the appropriate features and models to use remain unclear. We investigated the best features available from the PPG and ECG for BP estimation using both linear and non-linear machine learning models. We conducted a clinical study in which changes in BP (DeltaBP) were induced by an infusion of phenylephrine in 30 healthy volunteers (53.8% female, 28.0 (9.0) years old). We extracted a large and diverse set of features from both the PPG and the ECG and assessed their individual importance for estimating DeltaBP through Shapley additive explanation values and a ranking coefficient. We trained, tuned, and evaluated linear (ordinary least squares, OLS) and non-linear (random forest, RF) machine learning models to estimate DeltaBP in a nested leave-one-subject-out cross-validation framework. We reported the results as correlation coefficient (rho _p), root mean squared error (RMSE), and mean absolute error (MAE). The non-linear RF model significantly (p<0.05) outperformed the linear OLS model using both the PPG and the ECG signals across all performance metrics. Estimating DeltaSBP using the PPG alone (rho _p = 0.86 (0.23), RMSE = 5.66 (4.76) mmHg, MAE = 4.86 (4.29) mmHg) performed significantly better than using the ECG alone (rho _p = 0.69 (0.45), RMSE = 6.79 (4.76) mmHg, MAE = 5.28 (4.57) mmHg), all p < 0.001. The highest ranking features from the PPG largely modelled increasing reflected wave interference driven by changes in arterial stiffness. This finding was supported by changes observed in the PPG waveform in response to the phenylephrine infusion. However, a large number of features were required for accurate BP estimation, highlighting the high complexity of the problem. We conclude that the PPG alone may be further explored as a potential single source, cuffless, blood pressure estimator. The use of the ECG alone is not justified. Non-linear models may perform better as they are able to incorporate interactions between feature values and demographics. However, demographics may not adequately account for the unique and individualised relationship between the extracted features and BP.

Yield prediction through integration of genetic, environment, and management data through deep learning.

Accurate prediction of the phenotypic outcomes produced by different combinations of genotypes, environments, and management interventions remains a key goal in biology with direct applications to agriculture, research, and conservation. The past decades have seen an expansion of new methods applied toward this goal. Here we predict maize yield using deep neural networks, compare the efficacy of 2 model development methods, and contextualize model performance using conventional linear and machine learning models. We examine the usefulness of incorporating interactions between disparate data types. We find deep learning and best linear unbiased predictor (BLUP) models with interactions had the best overall performance. BLUP models achieved the lowest average error, but deep learning models performed more consistently with similar average error. Optimizing deep neural network submodules for each data type improved model performance relative to optimizing the whole model for all data types at once. Examining the effect of interactions in the best-performing model revealed that including interactions altered the model's sensitivity to weather and management features, including a reduction of the importance scores for timepoints expected to have a limited physiological basis for influencing yield-those at the extreme end of the season, nearly 200 days post planting. Based on these results, deep learning provides a promising avenue for the phenotypic prediction of complex traits in complex environments and a potential mechanism to better understand the influence of environmental and genetic factors.