Independent Test Data Research Articles

Predicting Limit-Setting Behavior of Gamblers Using Machine Learning Algorithms: A Real-World Study of Norwegian Gamblers Using Account Data

Player protection and harm minimization have become increasingly important in the gambling industry along with the promotion of responsible gambling (RG). Among the most widespread RG tools that gaming operators provide are limit-setting tools that help players limit the amount of time and/or money they spend gambling. Research suggests that limit-setting significantly reduces the amount of money that players spend. If limit-setting is to be encouraged as a way of facilitating responsible gambling, it is important to know what variables are important in getting individuals to set and change limits in the first place. In the present study, 33 variables assessing the player behavior among Norsk Tipping clientele (N = 70,789) from January to March 2017 were computed. The 33 variables which reflect the players’ behavior were then used to predict the likelihood of gamblers changing their monetary limit between April and June 2017. The 70,789 players were randomly split into a training dataset of 56,532 and an evaluation set of 14,157 players (corresponding to an 80/20 split). The results demonstrated that it is possible to predict future limit-setting based on player behavior. The random forest algorithm appeared to predict limit-changing behavior much better than the other algorithms. However, on the independent test data, the random forest algorithm’s accuracy dropped significantly. The best performance on the test data along with a small decrease in accuracy in comparison to the training data was delivered by the gradient boost machine learning algorithm. The most important variables predicting future limit-setting using the gradient boost machine algorithm were players receiving feedback that they had reached 80% of their personal monthly global loss limit, personal monthly loss limit, the amount bet, theoretical loss, and whether the players had increased their limits in the past. With the help of predictive analytics, players with a high likelihood of changing their limits can be proactively approached.

Rapid classification of group B Streptococcus serotypes based on matrix-assisted laser desorption ionization-time of flight mass spectrometry and machine learning techniques

BackgroundGroup B streptococcus (GBS) is an important pathogen that is responsible for invasive infections, including sepsis and meningitis. GBS serotyping is an essential means for the investigation of possible infection outbreaks and can identify possible sources of infection. Although it is possible to determine GBS serotypes by either immuno-serotyping or geno-serotyping, both traditional methods are time-consuming and labor-intensive. In recent years, the matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) has been reported as an effective tool for the determination of GBS serotypes in a more rapid and accurate manner. Thus, this work aims to investigate GBS serotypes by incorporating machine learning techniques with MALDI-TOF MS to carry out the identification.ResultsIn this study, a total of 787 GBS isolates, obtained from three research and teaching hospitals, were analyzed by MALDI-TOF MS, and the serotype of the GBS was determined by a geno-serotyping experiment. The peaks of mass-to-charge ratios were regarded as the attributes to characterize the various serotypes of GBS. Machine learning algorithms, such as support vector machine (SVM) and random forest (RF), were then used to construct predictive models for the five different serotypes (Types Ia, Ib, III, V, and VI). After optimization of feature selection and model generation based on training datasets, the accuracies of the selected models attained 54.9–87.1% for various serotypes based on independent testing data. Specifically, for the major serotypes, namely type III and type VI, the accuracies were 73.9 and 70.4%, respectively.ConclusionThe proposed models have been adopted to implement a web-based tool (GBSTyper), which is now freely accessible at http://csb.cse.yzu.edu.tw/GBSTyper/, for providing efficient and effective detection of GBS serotypes based on a MALDI-TOF MS spectrum. Overall, this work has demonstrated that the combination of MALDI-TOF MS and machine intelligence could provide a practical means of clinical pathogen testing.

Prediction of Suitable Habitat for Lycophytes and Ferns in Northeast China: A Case Study on Athyrium brevifrons

Suitable habitat is vital for the survival and restoration of a species. Understanding the suitable habitat range for lycophytes and ferns is prerequisite for effective species resource conservation and recovery efforts. In this study, we took Athyrium brevifrons as an example, predicted its suitable habitat using a Maxent model with 67 occurrence data and nine environmental variables in Northeast China. The area under the curve (AUC) value of independent test data, as well as the comparison with specimen county areal distribution of A. brevifrons exhibited excellent predictive performance. The type of environmental variables showed that precipitation contributed the most to the distribution prediction, followed by temperature and topography. Percentage contribution and permutation importance both indicated that precipitation of driest quarter (Bio17) was the key factor in determining the natural distribution of A. brevifrons, the reason could be proved by the fern gametophyte biology. The analysis of high habitat suitability areas also showed the habitat preference of A. brevifrons: comparatively more precipitation and less fluctuation in the driest quarter. Changbai Mountains, covering almost all the high and medium habitat suitability areas, provide the best ecological conditions for the survival of A. brevifrons, and should be considered as priority areas for protection and restoration of the wild resource. The potential habitat suitability distribution map could provide a reference for the sustainable development and utilisation of A. brevifrons resource, and Maxent modelling could be valuable for conservation management planning for lycophytes and ferns in Northeast China.

Three-dimensional Deep Convolutional Neural Networks for Automated Myocardial Scar Quantification in Hypertrophic Cardiomyopathy: A Multicenter Multivendor Study.

Background Cardiac MRI late gadolinium enhancement (LGE) scar volume is an important marker for outcome prediction in patients with hypertrophic cardiomyopathy (HCM); however, its clinical application is hindered by a lack of measurement standardization. Purpose To develop and evaluate a three-dimensional (3D) convolutional neural network (CNN)-based method for automated LGE scar quantification in patients with HCM. Materials and Methods We retrospectively identified LGE MRI data in a multicenter (n = 7) and multivendor (n = 3) HCM study obtained between November 2001 and November 2011. A deep 3D CNN based on U-Net architecture was used for LGE scar quantification. Independent CNN training and testing data sets were maintained with a 4:1 ratio. Stacks of short-axis MRI slices were split into overlapping substacks that were segmented and then merged into one volume. The 3D CNN per-site and per-vendor performances were evaluated with respect to manual scar quantification performed in a core laboratory setting using Dice similarity coefficient (DSC), Pearson correlation, and Bland-Altman analyses. Furthermore, the performance of 3D CNN was compared with that of two-dimensional (2D) CNN. Results This study included 1073 patients with HCM (733 men; mean age, 49 years ± 17 [standard deviation]). The 3D CNN-based quantification was fast (0.15 second per image) and demonstrated excellent correlation with manual scar volume quantification (r = 0.88, P < .001) and ratio of scar volume to total left ventricle myocardial volume (%LGE) (r = 0.91, P < .001). The 3D CNN-based quantification strongly correlated with manual quantification of scar volume (r = 0.82-0.99, P < .001) and %LGE (r = 0.90-0.97, P < .001) for all sites and vendors. The 3D CNN identified patients with a large scar burden (>15%) with 98% accuracy (202 of 207) (95% confidence interval [CI]: 95%, 99%). When compared with 3D CNN, 2D CNN underestimated scar volume (r = 0.85, P < .001) and %LGE (r = 0.83, P < .001). The DSC of 3D CNN segmentation was comparable among different vendors (P = .07) and higher than that of 2D CNN (DSC, 0.54 ± 0.26 vs 0.48 ± 0.29; P = .02). Conclusion In the hypertrophic cardiomyopathy population, a three-dimensional convolutional neural network enables fast and accurate quantification of myocardial scar volume, outperforms a two-dimensional convolutional neural network, and demonstrates comparable performance across different vendors. © RSNA, 2019 Online supplemental material is available for this article.

Application of proteometric approach for identification of functional mutant sites to improve the binding affinity of anticancer biologic trastuzumab with its antigen human epidermal growth factor receptor 2.

The aim of the present study was to develop a linear regression model aiding to a quick scan of the most important sites for mutation of an anticancer biologic trastuzumab. The important sites identified on trastuzumab can be used to carry out site-directed mutagenesis to improve the binding affinity of the drug towards its antigen, human epidermal growth factor receptor 2 (HER2). This will lead to low dosage requirement of the drug for treating cancer patients, which in turn help to cut the cost and combat development of resistance. A quantitative structure-activity relationship (QSAR) model was built by multiple linear regressions using genetic algorithm-based feature selection (GA-MLR) method using 48 dependent variables (dissociation constant Kd ) and 226 independent variables (theoretical descriptors generated using a proteometrics approach). The final QSAR model selected in the study was more on the basis of ability to predict accurately independent test data and generalization ability of the model rather than mere statistical significance of the model. With combined analysis of descriptors presented in final QSAR model and most frequent descriptors pooled from all solution models, it was demonstrated that the modeling procedure was able to bring on the factors important for antigen-antibody interactions with an example of HER2-trastuzumab interaction reported in previous experimental studies. This paper will allow the prediction of the most preferable site to mutate for improving the binding affinity of trastuzumab with HER2 and also will be helpful in selecting most preferable amino acids to substitute in the selected site for mutations. This is the novel report on proteometrics approach with autocorrelation formalism for antibody engineering, which can be extended to other antibody-antigen pairs.

Modelling seedling development using thermal effectiveness and photosynthetically active radiation

Seedling quality is a prerequisite for successful field performance and therefore influences crop yields. Temperature and illumination are two major factors affecting seedling quality during nursery propagation. Suboptimal temperature or light of nurseries generally result in leggy or weak seedlings and great economic loss. However, production of healthy seedlings is challenging due to the lack of knowledge in systemic management of nursery environments. In this study, we have established simulation models to predict how temperature and illumination coordinately influence the growth of tomato and cabbage seedlings. Specifically, correlation between seedling quality characteristics (root-shoot ratio, G value (growth function: defined as ratio of whole plant dry weight to days of seedling), healthy indexes) and TEP (thermal effectiveness and photosynthetically active radiation) were explored to establish the models, which were validated with independent test data. Our results suggested that the curve of healthy index 1 (HI1) and TEP fitted well with high coefficient of determination (R2) in both species, indicating that the model is highly reliable. The HI1 simulation models for tomato and cabbage are HI1=0.0009e0.0308TEP–0.0015 and HI1 = 0.0003e0.0671TEP–0.0003, respectively, which can be used for predicting vigors of tomato and cabbage seedlings grown under different temperature and light conditions.

Predicting plant available phosphorus using infrared spectroscopy with consideration for future mobile sensing applications in precision farming

Phosphorus (P) fertilisation recommendations rely primarily on soil content of plant available P (Pavl) that vary spatially within farm fields. Spatially optimized P fertilisation for precision farming requires reliable, rapid and non-invasive Pavl determination. This laboratory study aimed to test and to compare visible-near infrared (Vis–NIR) and mid-infrared (MIR) spectroscopy for Pavl prediction with emphasis on future application in precision agriculture. After calibration with the conventional calcium acetate lactate (CAL) extraction method, limitations of Vis–NIRS and MIRS to predict Pavl were evaluated in loess topsoil samples from different fields at six localities. Overall calibration with 477 (Vis–NIRS) and 586 (MIRS) samples yielded satisfactory model performance (R2 0.70 and 0.72; RPD 1.8 and 1.9, respectively). Local Vis–NIRS models yielded better results with R2 up to 0.93 and RPD up to 3.8. For MIRS, results were comparable. However, an overall model to predict Pavl on independent test data partly failed. Sampling date, pre-crop harvest residues and fertilising regime affected model transferability. Varying transferability could partly be explained after deriving the cellulose absorption index from the Vis–NIR spectra. In 62 (Vis–NIRS) and 67% (MIRS) of all samples, prediction matched the correct Pavl content class. Rapid discrimination between high, optimal and low P classes could be carried out on many samples from single fields thus marking an improvement over the common practice. However, Pavl determination by means of IR spectroscopy is not yet satisfactory for determination of precision fertilizer dosage. For introduction into agricultural practice, a standardized sampling protocol is recommended to help achieve reliable spectroscopic Pavl prediction.

A multi-source data integration approach reveals novel associations between metabolites and renal outcomes in the German Chronic Kidney Disease study

Omics data facilitate the gain of novel insights into the pathophysiology of diseases and, consequently, their diagnosis, treatment, and prevention. To this end, omics data are integrated with other data types, e.g., clinical, phenotypic, and demographic parameters of categorical or continuous nature. We exemplify this data integration issue for a chronic kidney disease (CKD) study, comprising complex clinical, demographic, and one-dimensional 1H nuclear magnetic resonance metabolic variables. Routine analysis screens for associations of single metabolic features with clinical parameters while accounting for confounders typically chosen by expert knowledge. This knowledge can be incomplete or unavailable. We introduce a framework for data integration that intrinsically adjusts for confounding variables. We give its mathematical and algorithmic foundation, provide a state-of-the-art implementation, and evaluate its performance by sanity checks and predictive performance assessment on independent test data. Particularly, we show that discovered associations remain significant after variable adjustment based on expert knowledge. In contrast, we illustrate that associations discovered in routine univariate screening approaches can be biased by incorrect or incomplete expert knowledge. Our data integration approach reveals important associations between CKD comorbidities and metabolites, including novel associations of the plasma metabolite trimethylamine-N-oxide with cardiac arrhythmia and infarction in CKD stage 3 patients.

High-resolution mapping of forest vulnerability to wind for disturbance-aware forestry

Windstorms cause major disturbances in European forests. Forest management can play a key role in making forests more persistent to disturbances. However, better information is needed to support decision making that effectively accounts for wind disturbances. Here we show how empirical probability models of wind damage, combined with existing spatial data sets, can be used to provide fine-scale spatial information about disturbance probability over large areas. First, we created stand-level damage probability models using wind damage observations within 5-year time window in national forest inventory data (NFI). Model predictors described forest characteristics, forest management history, 10-year return-rate of maximum wind speed, and soil, site and climate conditions. We tested three different methods for creating the damage probability models – generalized linear models (GLM), generalized additive models (GAM) and boosted regression trees (BRT). Then, damage probability maps were calculated by combining the models with GIS data sets representing the model predictors. Finally, we demonstrated the predictive performance of the damage probability maps with a large, independent test data of over 33,000 NFI plots, which shows that the maps are able to identify vulnerable forests also in new wind damage events, with area under curve value (AUC) > 0.7. Use of the more complex methods (GAM and BRT) was not found to improve the performance of the map compared to GLM, and therefore we prefer using the simpler GLM method that can be more easily interpreted. The map allows identification of vulnerable forest areas in high spatial resolution (16 m × 16 m), making it useful in assessing the vulnerability of individual forest stands when making management decisions. The map is also a powerful tool for communicating disturbance risks to forest owners and managers and it has the potential to steer forest management practices to a more disturbance-aware direction. Our study showed that in spite of the inherent stochasticity of the wind and damage phenomena at all spatial scales, it can be modelled with good accuracy across large spatial scales when existing ground and earth observation data sources are combined smartly. With improving data quality and availability, map-based risk assessments can be extended to other regions and other disturbance types.

Prediction of complication related death after radical cystectomy for bladder cancer with machine learning methodology

Purpose: To create a pre-operatively usable tool to identify patients at high risk of early death (within 90 days post-operatively) after radical cystectomy and to assess potential risk factors for post-operative and surgery related mortality.Materials and methods: Material consists of 1099 consecutive radical cystectomy (RC) patients operated at 16 different hospitals in Finland 2005–2014. Machine learning methodology was utilized. For model building and testing, the data was randomly divided into training data (n = 733, 66.7%) and independent testing data (n = 366, 33.3%). To predict the risk of early death after RC from baseline variables, a binary classifier was constructed using logistic regression with lasso regularization. Finally, a user-friendly risk table was constructed for practical use.Results: The model resulted in an area under the receiver operating characteristic curve (AUROC) of 0.73 (95% CI = 0.59–0.87). The strongest risk factors were: American Society of Anesthesiologists physical status classification (ASA), congestive heart failure (CHF), age adjusted Charlson comorbidity index (ACCI) and chronic pulmonary disease.Conclusion: This study with a novel methodological approach adds CHF and chronic pulmonary disease to previously known independent prognostic risk factors for early death after RC. Importantly, the risk prediction tool uses purely pre-operative data and can be used before surgery.

Age and Sex Estimation Using Artificial Intelligence From Standard 12-Lead ECGs.

Sex and age have long been known to affect the ECG. Several biologic variables and anatomic factors may contribute to sex and age-related differences on the ECG. We hypothesized that a convolutional neural network (CNN) could be trained through a process called deep learning to predict a person's age and self-reported sex using only 12-lead ECG signals. We further hypothesized that discrepancies between CNN-predicted age and chronological age may serve as a physiological measure of health. We trained CNNs using 10-second samples of 12-lead ECG signals from 499 727 patients to predict sex and age. The networks were tested on a separate cohort of 275 056 patients. Subsequently, 100 randomly selected patients with multiple ECGs over the course of decades were identified to assess within-individual accuracy of CNN age estimation. Of 275 056 patients tested, 52% were males and mean age was 58.6±16.2 years. For sex classification, the model obtained 90.4% classification accuracy with an area under the curve of 0.97 in the independent test data. Age was estimated as a continuous variable with an average error of 6.9±5.6 years (R-squared =0.7). Among 100 patients with multiple ECGs over the course of at least 2 decades of life, most patients (51%) had an average error between real age and CNN-predicted age of <7 years. Major factors seen among patients with a CNN-predicted age that exceeded chronologic age by >7 years included: low ejection fraction, hypertension, and coronary disease (P<0.01). In the 27% of patients where correlation was >0.8 between CNN-predicted and chronologic age, no incident events occurred over follow-up (33±12 years). Applying artificial intelligence to the ECG allows prediction of patient sex and estimation of age. The ability of an artificial intelligence algorithm to determine physiological age, with further validation, may serve as a measure of overall health.

RF_Purify: a novel tool for comprehensive analysis of tumor-purity in methylation array data based on random forest regression

BackgroundWith the advent of array-based techniques to measure methylation levels in primary tumor samples, systematic investigations of methylomes have widely been performed on a large number of tumor entities. Most of these approaches are not based on measuring individual cell methylation but rather the bulk tumor sample DNA, which contains a mixture of tumor cells, infiltrating immune cells and other stromal components. This raises questions about the purity of a certain tumor sample, given the varying degrees of stromal infiltration in different entities. Previous methods to infer tumor purity require or are based on the use of matching control samples which are rarely available. Here we present a novel, reference free method to quantify tumor purity, based on two Random Forest classifiers, which were trained on ABSOLUTE as well as ESTIMATE purity values from TCGA tumor samples. We subsequently apply this method to a previously published, large dataset of brain tumors, proving that these models perform well in datasets that have not been characterized with respect to tumor purity .ResultsUsing two gold standard methods to infer purity – the ABSOLUTE score based on whole genome sequencing data and the ESTIMATE score based on gene expression data- we have optimized Random Forest classifiers to predict tumor purity in entities that were contained in the TCGA project. We validated these classifiers using an independent test data set and cross-compared it to other methods which have been applied to the TCGA datasets (such as ESTIMATE and LUMP).Using Illumina methylation array data of brain tumor entities (as published in Capper et al. (Nature 555:469-474,2018)) we applied this model to estimate tumor purity and find that subgroups of brain tumors display substantial differences in tumor purity.ConclusionsRandom forest- based tumor purity prediction is a well suited tool to extrapolate gold standard measures of purity to novel methylation array datasets. In contrast to other available methylation based tumor purity estimation methods, our classifiers do not need a priori knowledge about the tumor entity or matching control tissue to predict tumor purity.

Computational Prediction of Ubiquitination Proteins Using Evolutionary Profiles and Functional Domain Annotation

Background: Ubiquitination, as a post-translational modification, is a crucial biological process in cell signaling, apoptosis, and localization. Identification of ubiquitination proteins is of fundamental importance for understanding the molecular mechanisms in biological systems and diseases. Although high-throughput experimental studies using mass spectrometry have identified many ubiquitination proteins and ubiquitination sites, the vast majority of ubiquitination proteins remain undiscovered, even in well-studied model organisms.Objective: To reduce experimental costs, computational methods have been introduced to predict ubiquitination sites, but the accuracy is unsatisfactory. If it can be predicted whether a protein can be ubiquitinated or not, it will help in predicting ubiquitination sites. However, all the computational methods so far can only predict ubiquitination sites.Methods: In this study, the first computational method for predicting ubiquitination proteins without relying on ubiquitination site prediction has been developed. The method extracts features from sequence conservation information through a grey system model, as well as functional domain annotation and subcellular localization.Results: Together with the feature analysis and application of the relief feature selection algorithm, the results of 5-fold cross-validation on three datasets achieved a high accuracy of 90.13%, with Matthew’s correlation coefficient of 80.34%. The predicted results on an independent test data achieved 87.71% as accuracy and 75.43% of Matthew’s correlation coefficient, better than the prediction from the best ubiquitination site prediction tool available.Conclusion: Our study may guide experimental design and provide useful insights for studying the mechanisms and modulation of ubiquitination pathways. The code is available at: https://github.com/Chunhuixu/UBIPredic_QWRCHX

Estimating aboveground biomass in subtropical forests of China by integrating multisource remote sensing and ground data

The biomass of the subtropical forests of China is an important component of the global carbon cycle. Recently, several above ground biomass (AGB) maps have been produced using a variety of approaches to assess the carbon stock of the subtropical forest in China. However, due to the lack of reliable ground observations and the limitations of AGB mapping methods at regional scales, estimates of the spatial distribution of AGB vary greatly, leading to large uncertainties in the carbon stock estimations. In this study, we produced a new 1-km spatial resolution AGB map by synthesizing an unprecedented number of ground AGB observations from published studies, and developed an AGB mapping method using a combination of ground observations, MODIS data, forest cover/gain/loss maps based on Landsat, GLAS forest canopy height, and climatic and terrain data. In addition, we validated our estimates using independent testing data and compared our estimates with three previous AGB maps. The results indicate that the total AGB stock in the subtropical forest of China is (266 ± 9.1) × 106 Mg, with an average AGB of 123.2 Mg/ha. Based on sixteen explanatory variables, our ensemble mean model explains 75% of the variance in forest AGB, with an RMSE of 45.5 Mg/ha. Comparison using all observation data shows that our map has a significantly lower RMSE and bias than previous maps, where the RMSE and bias tended to vary with forest type. This study not only improved the accuracy of AGB estimation for the subtropical forests but also highlighted the importance of forest type for regional AGB estimation.

Potato Yield Prediction Using Machine Learning Techniques and Sentinel 2 Data

Traditional potato growth models evidence certain limitations, such as the cost of obtaining the input data required to run the models, the lack of spatial information in some instances, or the actual quality of input data. In order to address these issues, we develop a model to predict potato yield using satellite remote sensing. In an effort to offer a good predictive model that improves the state of the art on potato precision agriculture, we use images from the twin Sentinel 2 satellites (European Space Agency—Copernicus Programme) over three growing seasons, applying different machine learning models. First, we fitted nine machine learning algorithms with various pre-processing scenarios using variables from July, August and September based on the red, red-edge and infra-red bands of the spectrum. Second, we selected the best performing models and evaluated them against independent test data. Finally, we repeated the previous two steps using only variables corresponding to July and August. Our results showed that the feature selection step proved vital during data pre-processing in order to reduce multicollinearity among predictors. The Regression Quantile Lasso model (11.67% Root Mean Square Error, RMSE; R2 = 0.88 and 9.18% Mean Absolute Error, MAE) and Leap Backwards model (10.94% RMSE, R2 = 0.89 and 8.95% MAE) performed better when predictors with a correlation coefficient &gt; 0.5 were removed from the dataset. In contrast, the Support Vector Machine Radial (svmRadial) performed better with no feature selection method (11.7% RMSE, R2 = 0.93 and 8.64% MAE). In addition, we used a random forest model to predict potato yields in Castilla y León (Spain) 1–2 months prior to harvest, and obtained satisfactory results (11.16% RMSE, R2 = 0.89 and 8.71% MAE). These results demonstrate the suitability of our models to predict potato yields in the region studied.

Quality of local scale surface weather analogs over the north-west Himalaya (NWH), India

An analog ensemble system was developed for the realisation of local-scale surface meteorological variables for independent test data (test data) at six stations over the north-west Himalaya (NWH), India. Extreme values (the maximum value and the minimum value) and the mean value in 10 analog days (the analog mean) and the climatological mean of each surface meteorological variable were compared with its corresponding observed values on the same day (d0, lead time 0 hour (h)), d1 (d0 + 1, lead time 24 h), d2 (d0 + 2, lead time 48 h) and d3 (d0 + 3, lead time 72 h) of test data. Pearson correlation coefficients (CCs), Mean Absolute Differences (MADs) and Root Mean Square Differences (RMSDs) of the extreme values in analog days, and the analog mean and climatological mean of each meteorological variable on d0 with its corresponding observed values on d0, d1, d2 and d3 of test data were computed at six stations over the NWH. CCs of extreme values in analog days and the analog mean of each meteorological variable on d0 with its observed values on d0, d1, d2 and d3 were found to be higher than the CCs of the climatological mean of each meteorological variable on d0 with its observed values on d0, d1, d2 and d3. MADs (RMSDs) of extreme values in analog days and the analog mean of each meteorological variable on d0 with its observed values on d0, d1, d2 and d3 were found to be lesser than the MADs (RMSDs) of the climatological mean of each meteorological variable on d0 with its observed values on d0, d1, d2 and d3. However, the MADs (RMSDs) of the extreme values of each meteorological variable in analog days were found to be higher than the MADs (RMSDs) of its analog mean. Results show that the analog mean of each meteorological variable holds better predictive skill than the extreme values in analog days and its climatological mean. MADs (RMSDs) of different surface meteorological variables in surface weather analogs comparable to Mean Absolute Errors (MAEs) and Root Mean Square Errors (RMSEs) for their prediction with the help of different types of weather forecast models show that the surface weather analogs hold good promise for the local-scale prediction of surface meteorological variables over the NWH.

Optimized machine learning methods for prediction of cognitive outcome in Parkinson's disease

BackgroundGiven the increasing recognition of the significance of non-motor symptoms in Parkinson's disease, we investigate the optimal use of machine learning methods for the prediction of the Montreal Cognitive Assessment (MoCA) score at year 4 from longitudinal data obtained at years 0 and 1. MethodsWe selected n = 184 PD subjects from the Parkinson's Progressive Marker Initiative (PPMI) database (93 features). A range of robust predictor algorithms (accompanied with automated machine learning hyperparameter tuning) and feature subset selector algorithms (FSSAs) were selected. We utilized 65%, 5% and 30% of patients in each arrangement for training, training validation and final testing respectively (10 randomized arrangements). For further testing, we enrolled 308 additional patients. ResultsFirst, we employed 10 predictor algorithms, provided with all 93 features; an error of 1.83 ± 0.13 was obtained by LASSOLAR (Least Absolute Shrinkage and Selection Operator - Least Angle Regression). Subsequently, we used feature subset selection followed by predictor algorithms. GA (Genetic Algorithm) selected 18 features; subsequently LOLIMOT (Local Linear Model Trees) reached an error of 1.70 ± 0.10. DE (Differential evolution) also selected 18 features and coupled with Thiel-Sen regression arrived at a similar performance. NSGAII (Non-dominated sorting genetic algorithm) yielded the best performance: it selected six vital features, which combined with LOLIMOT reached an error of 1.68 ± 0.12. Finally, using this last approach on independent test data, we reached an error of 1.65. ConclusionBy employing appropriate optimization tools (including automated hyperparameter tuning), it is possible to improve prediction of cognitive outcome. Overall, we conclude that optimal utilization of FSSAs and predictor algorithms can produce very good prediction of cognitive outcome in PD patients.

A New Plagioclase-Liquid Hygrometer Specific to Trachytic Systems

We present a new empirical plagioclase-liquid hygrometer for estimating the amount of H2O dissolved in trachytic magmas. The hygrometer is based on the exchange reaction of anorthite between plagioclase and liquid, and is calibrated using crystallization experiments where the concentration of H2O in quenched glasses has been accurately determined based on Fourier Transform Infrared Spectroscopy (FTIR) analysis. The multiple linear regression of plagioclase-liquid cation fractions and components from experimental data obtained at 150–202 MPa, 850–1020 °C, 1.17–7.57 wt. % H2O and ΔNNO + 2.5 buffer, yields to a highly accurate model with uncertainty of only ±0.29 wt. % H2O. The model reliability has been demonstrated using an independent test data set consisting of crystallization experiments from the literature and thermodynamically derived compositions. The fairly good convergence between our model calibration and the test data set excludes systematic H2O overestimates or underestimates caused by miscalibration and data overfitting. The plagioclase-liquid hygrometer from this study has been applied to trachyandesitic (latitic) and trachytic products erupted over the last 1000 years at the La Fossa cone of Vulcano Island (Aeolian Islands, Southern Italy). Results from calculations indicate that the concentration of H2O in the latitic and trachytic melts is comprised between ~2.5 and ~3.5 wt. %. These values are in good agreement with data from melt inclusions and, overall, testify to low-pressure, open-system differentiation of trachytic magmas under strong degassing conditions.

Using Deep Learning to Estimate Tropical Cyclone Intensity from Satellite Passive Microwave Imagery

AbstractA deep learning convolutional neural network model is used to explore the possibilities of estimating tropical cyclone (TC) intensity from satellite images in the 37- and 85–92-GHz bands. The model, called “DeepMicroNet,” has unique properties such as a probabilistic output, the ability to operate from partial scans, and resiliency to imprecise TC center fixes. The 85–92-GHz band is the more influential data source in the model, with 37 GHz adding a marginal benefit. Training the model on global best track intensities produces model estimates precise enough to replicate known best track intensity biases when compared to aircraft reconnaissance observations. Model root-mean-square error (RMSE) is 14.3 kt (1 kt ≈ 0.5144 m s−1) compared to two years of independent best track records, but this improves to an RMSE of 10.6 kt when compared to the higher-standard aircraft reconnaissance-aided best track dataset, and to 9.6 kt compared to the reconnaissance-aided best track when using the higher-resolution TRMM TMI andAquaAMSR-E microwave observations only. A shortage of training and independent testing data for category 5 TCs leaves the results at this intensity range inconclusive. Based on this initial study, the application of deep learning to TC intensity analysis holds tremendous promise for further development with more advanced methodologies and expanded training datasets.

Cognitive self-regulation influences pain-related physiology.

Cognitive self-regulation can shape pain experience, but its effects on autonomic responses to painful events are unclear. In this study, participants (N = 41) deployed a cognitive strategy based on reappraisal and imagination to regulate pain up or down on different trials while skin conductance responses (SCRs) and electrocardiogram activity were recorded. Using a machine learning approach, we first developed stimulus-locked SCR and electrocardiogram physiological markers predictive of pain ratings. The physiological markers demonstrated high sensitivity and moderate specificity in predicting pain across 2 data sets, including an independent test data set (N = 84). When we tested the markers on the cognitive self-regulation data, we found that cognitive self-regulation had significant impacts on both pain ratings and pain-related physiology in accordance with regulatory goals. These findings suggest that self-regulation can impact autonomic nervous system responses to painful stimuli and provide pain-related autonomic profiles for future studies.