Nonparametric Statistical Model Research Articles

Development of a quantitative prediction algorithm for human cord blood-derived CD34+ hematopoietic stem-progenitor cells using parametric and non-parametric machine learning models

The transplantation of CD34+ hematopoietic stem-progenitor cells (HSPCs) derived from cord blood serves as the standard treatment for selected hematological, oncological, metabolic, and immunodeficiency disorders, of which the dose is pivotal to the clinical outcome. Based on numerous maternal and neonatal parameters, we evaluated the predictive power of mathematical pipelines to the proportion of CD34+ cells in the final cryopreserved cord blood product adopting both parametric and non-parametric algorithms. Twenty-four predictor variables associated with the cord blood processing of 802 processed cord blood units randomly sampled in 2020–2022 were retrieved and analyzed. Prediction models were developed by adopting the parametric (multivariate linear regression) and non-parametric (random forest and back propagation neural network) statistical models to investigate the data patterns for determining the single outcome (i.e., the proportion of CD34+ cells). The multivariate linear regression model produced the lowest root-mean-square deviation (0.0982). However, the model created by the back propagation neural network produced the highest median absolute deviation (0.0689) and predictive power (56.99%) in comparison to the random forest and multivariate linear regression. The predictive model depending on a combination of continuous and discrete maternal with neonatal parameters associated with cord blood processing can predict the CD34+ dose in the final product for clinical utilization. The back propagation neural network algorithm produces a model with the highest predictive power which can be widely applied to assisting cell banks for optimal cord blood unit selection to ensure the highest chance of transplantation success.

Seasonality in synthetic average wind speed

There is a growing demand for computer-generated realistic high-fidelity wind speed data for various applications in the wind industry. Such data should capture the non-stationary dynamics of real-world wind time series, as well as be consistent with the statistical descriptors – the probability density function and power spectral density – of the observed wind speed. However, complying with the statistical descriptors is not a guarantee that the seasonality will be correctly reproduced in synthetic data. The seasonality, characterized by the average diurnal and seasonal variations, is driven by the periodicities embedded in diurnal and annual harmonic series respectively. Those periodicities are determined by the long-term orbital forcing components, which establish the insolation for a given latitude and longitude. We show that average diurnal and seasonal variations can be visualized as the output of comb filters, whose fundamental frequencies match the diurnal and annual fundamental frequency respectively. The aforementioned theoretical findings are readily reproduced in synthetic wind speed, generated by a non-parametric data-driven statistical model, based on the phase-randomized Fourier transform. The model, tested on both 10-min and 1-min resolution real-world datasets, yields average non-stationarities in synthetic wind speed with the accuracy close to the computing precision.

The Importance of the Product “Tourism in Bullfighting Ranches” in Spain from the Perspective of the Breeders

Bullfighting tourism is a recent tourism modality that is in constant growth, and which consists of visits to bullfighting ranches with the aim of showing what life is like for this animal in its natural habitat. This activity represents a source of extra income for many breeders, but it has not yet been adequately promoted and publicized. This study analyzes the supply side of this type of tourism. For this purpose, a non-parametric statistical model has been used and results have been obtained which corroborate that the promotion of this type of tourism is scarce, and that the breeders began to offer this type of tourism to make known what life is like for the fighting bull in the countryside and to increase their income. Furthermore, in this study, it is fundamental to differentiate between bullfighting and bullfighting tourism so that this type of tourism can reach the greatest possible number of people and not be influenced by thoughts against bullfighting, as they are different activities.

Predicting and understanding residential water use with interpretable machine learning

Predicting residential water use is critical to efficiently manage urban water resource systems. Simultaneously, understanding the factors driving residential water use is required to plan for future urban change and achieve effective water resource management. Current approaches examining residential water use identify the drivers of household water use through parametric or non-parametric statistical approaches. Parametric approaches have high predictive errors and lack the ability to accurately capture interactions between features but allow for easy interpretation. Non-parametric approaches have lower predictive errors and can capture non-linear feature interactions but do not allow for easy interpretation. We use non-parametric statistical models of household water use and recent advances in interpretable machine learning to understand the drivers of residential water use. Specifically, we use post-hoc interpretability methods to examine how drivers of water use interact, focusing on environmental, demographic, physical housing, and utility policy factors. We find all four categories of factors are important for estimating water use with environmental and utility policy factors playing the largest role. Additionally, we identify non-linear interactions between many variables within and across these classes. We show this approach provides both high predictive accuracy and identification of complex water use factors, offering important insight for urban water management.

Enhancing Algorithm Selection through Comprehensive Performance Evaluation: Statistical Analysis of Stochastic Algorithms

Analyzing stochastic algorithms for comprehensive performance and comparison across diverse contexts is essential. By evaluating and adjusting algorithm effectiveness across a wide spectrum of test functions, including both classical benchmarks and CEC-C06 2019 conference functions, distinct patterns of performance emerge. In specific situations, underscoring the importance of choosing algorithms contextually. Additionally, researchers have encountered a critical issue by employing a statistical model randomly to determine significance values without conducting other studies to select a specific model for evaluating performance outcomes. To address this concern, this study employs rigorous statistical testing to underscore substantial performance variations between pairs of algorithms, thereby emphasizing the pivotal role of statistical significance in comparative analysis. It also yields valuable insights into the suitability of algorithms for various optimization challenges, providing professionals with information to make informed decisions. This is achieved by pinpointing algorithm pairs with favorable statistical distributions, facilitating practical algorithm selection. The study encompasses multiple nonparametric statistical hypothesis models, such as the Wilcoxon rank-sum test, single-factor analysis, and two-factor ANOVA tests. This thorough evaluation enhances our grasp of algorithm performance across various evaluation criteria. Notably, the research addresses discrepancies in previous statistical test findings in algorithm comparisons, enhancing result reliability in the later research. The results proved that there are differences in significance results, as seen in examples like Leo versus the FDO, the DA versus the WOA, and so on. It highlights the need to tailor test models to specific scenarios, as p-value outcomes differ among various tests within the same algorithm pair.

Molybdenum is not a risk factor for changes in serum testosterone

Molybdenum is a bio-essential trace nutrient, and elevated urinary molybdenum has been implicated as correlating with lower testosterone levels in men. To properly address the contention that molybdenum constitutes a potential health risk factor, we reviewed statistical research findings on molybdenum exposure and somatic levels in humans in a biological plausibility context, which is an indispensable component of reliable risk factor determination. Our analysis applied advanced nonparametric statistical modeling that expands linear regressions to include more potential variables and common confounders. The analysis shows that previously published negative molybdenum-testosterone associations no longer retain statistical relevance. In place of molybdenum as a stressor responsible for lowered testosterone levels, body mass index, age, related hormone variability, and underlying global downward trends in testosterone are shown to be the factors responsible for this effect.

Post-Operative Medium- and Long-Term Endocrine Outcomes in Patients with Non-Functioning Pituitary Adenomas-Machine Learning Analysis.

Post-operative endocrine outcomes in patients with non-functioning pituitary adenoma (NFPA) are variable. The aim of this study was to use machine learning (ML) models to better predict medium- and long-term post-operative hypopituitarism in patients with NFPAs. We included data from 383 patients who underwent surgery with or without radiotherapy for NFPAs, with a follow-up period between 6 months and 15 years. ML models, including k-nearest neighbour (KNN), support vector machine (SVM), and decision tree models, showed a superior ability to predict panhypopituitarism compared with non-parametric statistical modelling (mean accuracy: 0.89; mean AUC-ROC: 0.79), with SVM achieving the highest performance (mean accuracy: 0.94; mean AUC-ROC: 0.88). Pre-operative endocrine function was the strongest feature for predicting panhypopituitarism within 1 year post-operatively, while endocrine outcomes at 1 year post-operatively supported strong predictions of panhypopituitarism at 5 and 10 years post-operatively. Other features found to contribute to panhypopituitarism prediction were age, volume of tumour, and the use of radiotherapy. In conclusion, our study demonstrates that ML models show potential in predicting post-operative panhypopituitarism in the medium and long term in patients with NFPM. Future work will include incorporating additional, more granular data, including imaging and operative video data, across multiple centres.

Linear Mixed-Model Analysis Better Captures Subcomponents of Attention in a Small Sample Size of Persons With Aphasia.

Although there are several reports of attention deficits in aphasia, studies are typically limited to a single component within this complex domain. Furthermore, interpretation of results is affected by small sample size, intraindividual variability, task complexity, or nonparametric statistical models of performance comparison. The purpose of this study is to explore multiple subcomponents of attention in persons with aphasia (PWA) and compare findings and implications from various statistical methods-nonparametric, mixed analysis of variance (ANOVA), and linear mixed-effects model (LMEM)-when applied to a small sample size. Eleven PWA and nine age- and education-matched healthy controls (HCs) completed the computer-based Attention Network Test (ANT). ANT examines the effects of four types of warning cues (no, double, central, spatial) and two flanker conditions (congruent, incongruent) to provide an efficient way to assess the three subcomponents of attention (alerting, orienting, and executive control). Individual response time and accuracy data from each participant are considered for data analysis. Nonparametric analyses showed no significant differences between the groups on the three subcomponents of attention. Both mixed ANOVA and LMEM showed statistical significance on alerting effect in HCs, orienting effect in PWA, and executive control effect in both PWA and HCs. However, LMEM analyses additionally highlighted significant differences between the groups (PWA vs. HCs) for executive control effect, which were not evident on either ANOVA or nonparametric tests. By considering the random effect of participant ID, LMEM was able to show deficits in alerting and executive control ability in PWA when compared to HCs. LMEM accounts for the intraindividual variability based on individual response time performances instead of relying on measures of central tendencies.

Spectro-temporal acoustic elements of music interact in an integrated way to modulate emotional responses in pigs

Music is a complex stimulus, with various spectro-temporal acoustic elements determining one of the most important attributes of music, the ability to elicit emotions. Effects of various musical acoustic elements on emotions in non-human animals have not been studied with an integrated approach. However, this knowledge is important to design music to provide environmental enrichment for non-human species. Thirty-nine instrumental musical pieces were composed and used to determine effects of various acoustic parameters on emotional responses in farm pigs. Video recordings (n = 50) of pigs in the nursery phase (7–9 week old) were gathered and emotional responses induced by stimuli were evaluated with Qualitative Behavioral Assessment (QBA). Non-parametric statistical models (Generalized Additive Models, Decision Trees, Random Forests, and XGBoost) were applied and compared to evaluate relationships between acoustic parameters and pigs’ observed emotional responses. We concluded that musical structure affected emotional responses of pigs. The valence of modulated emotions depended on integrated and simultaneous interactions of various spectral and temporal structural components of music that can be readily modified. This new knowledge supports design of musical stimuli to be used as environmental enrichment for non-human animals.

Deep Gaussian processes and infinite neural networks for the analysis of EEG signals in Alzheimer’s diseases

Deep neural network models (DGPs) can be represented hierarchically by a sequential composition of layers. When the prior distribution over the weights and biases are independently identically distributed, there is an equivalence with Gaussian processes (GP) in the limit of an infinite net[1]work width. DGPs are non-parametric statistical models used to character[1]ize patterns of complex non-linear systems due to their flexibility, greater generalization capacity, and a natural way of making inferences about the parameters and states of the system. This article proposes a hierarchi[1]cal Bayesian structure to model the weights and biases of a deep neural network. We deduce a general formula to calculate the integrals of Gaussian processes with non-linear transfer densities and obtain a kernel to estimate the covariance functions. In the methodology, we conduct an empirical study analyzing an electroencephalogram (EEG) database for diagnosing Alzheimer’s disease. Additionally, the DGPs models are esti[1]mated and compared with the NN models for 5, 10, 50, 100, 500, and 1000 neurons in the hidden layer, considering two transfer functions: Recti[1]fied Linear Unit (ReLU) and hyperbolic Tangent (Tanh). The results show good performance in the classification of the signals. Finally, we use the mean square error as a goodness of fit measure to validate the proposed models, obtaining low estimation errors.

IDF21-0223 Diabetic foot ulcers’ impact on survival: an association study using the conditional inference tree approach

Background: Up to a third of people with diabetes worldwide will develop a diabetic foot ulcer (DFU) during their lifetime. DFUs precede 80% of limb amputations (LEA). At five years, mortality after diabetes-related LEA exceeds 70%, and the risk of death for people living with DFU is 2.5 times the risk for those without one. However, risk factors are still not fully understood. Aim: To study the association between DFU and mortality in adults with diabetes adjusted for modifiable and non-modifiable risk factors. Method: Cohort with 595 adults with diabetes who attended a Podiatry appointment at a Portuguese Diabetic Foot Outpatient Clinic, from January 2010 to March 2013, and followed up to 31 December 2020, or death. The conditional inference tree (CIT) was used to identify predictors of mortality. CIT is a nonparametric statistical model used for making classifications or predictions based on tree-structure recursive partitioning. The variables considered were: age, gender, diabetes duration, insulin therapy, glycemic control, living alone, retinopathy, physical impairment, visual impairment, nephropathy, lower limb revascularization, diabetic peripheral neuropathy (DPN), cardiovascular disease, previous DFU, previous LEA, hypolipidemic and anticoagulant agents. All the statistical analyses were carried out using the package partykit, in R software. Results: At baseline, diabetes median duration was 15 (IQ [inter-quartile] range 10-23) years, with the participants' median age of 67.3 (59.2-75.4) years. Participants were mostly male (57.1), with previous DFU (54.4%), with DPN (56.5%) and cardiovascular disease (83.4%); 10.7% had previous LEA, 46.4% had retinopathy, 19.3% nephropathy. During a median follow-up of 8.1 (IQ range 3.9-9.3) years, there were 318 (53.4%) deaths. The predictors of all-cause mortality identified by the CTI were previous DFU, age, and retinopathy. At the end of eight years of follow-up, the estimated survival probability (95% CI [confidence interval]) was 47.5 (35.6-63.4) % for people without a previous DFU, aged older than 65 years, and with retinopathy. For those without retinopathy, it was 73.7 (64.6-84.0) %. For people with a previous DFU and older than 65, the eight-year estimated survival probability was 25.0 (19.6-31.9) %, while for the younger ones, it was 56.9 (48.5-66.6) %. Retinopathy did not impact the survival probability for people with previous DFU. Discussion: In this cohort of adults with diabetes, age, previous DFU, and retinopathy were the risk factors with a significant association with mortality. For participants with previous DFU, the highest mortality was observed for those aged older than 65. However, for adults with no previous DFU, mortality was associated not only with being older than 65 years but also with retinopathy. Retinopathy and DFU, both frequent, preventable and markers of diabetes control, significantly affected survival after adjustment for age.

Application Statistical Models for Interpretation Toxicological Data

The results of bioassay on infusoria of 75 samples of bottom sediments from 6 water bodies of the Middle Volga region were analyzed using traditional nonparametric methods and statistical models Generalized linear mixed model (GLMM) and Cumulative link mixed model (CLMM). The ambiguity of the interpretation of the results of biotesting performed by nonparametric methods is due to the fact that the toxicological data often do not correspond to the normal distribution. The use of the GLMM and CLMM models allow analyze data that do not correspond to the normal distribution and made it possible to clarify the level of toxicity of a number of ambiguous samples, which, after processing by mathematical model algorithms, acquire the status of either exactly toxic or exactly non-toxic.

The cyclic interaction between daytime behavior and the sleep behavior of laboratory dogs

Sleep deprivation has been found to negatively affect an individual´s physical and psychological health. Sleep loss affects activity patterns, increases anxiety-like behaviors, decreases cognitive performance and is associated with depressive states. The activity/rest cycle of dogs has been investigated before, but little is known about the effects of sleep loss on the behavior of the species. Dogs are polyphasic sleepers, meaning the behavior is most observed at night, but bouts are also present during the day. However, sleep can vary with ecological and biological factors, such as age, sex, fitness, and even human presence. In this study, kennelled laboratory adult dogs’ sleep and diurnal behavior were recorded during 24-h, five-day assessment periods to investigate sleep quality and its effect on daily behavior. In total, 1560 h of data were analyzed, and sleep metrics and diurnal behavior were quantified. The relationship between sleeping patterns and behavior and the effect of age and sex were evaluated using non-parametric statistical tests and GLMM modelling. Dogs in our study slept substantially less than previously reported and presented a modified sleep architecture with fewer awakenings during the night and almost no sleep during the day. Sleep loss increased inactivity, decreased play and alert behaviors, while increased time spent eating during the day. Males appeared to be more affected by sleep fragmentation than females. Different age groups also experienced different effects of sleep loss. Overall, dogs appear to compensate for the lack of sleep during the night by remaining inactive during the day. With further investigations, the relationship between sleep loss and behavior has the potential to be used as a measure of animal welfare.

Flexible, non-parametric modeling using regularized neural networks

Non-parametric, additive models are able to capture complex data dependencies in a flexible, yet interpretable way. However, choosing the format of the additive components often requires non-trivial data exploration. Here, as an alternative, we propose PrAda-net, a one-hidden-layer neural network, trained with proximal gradient descent and adaptive lasso. PrAda-net automatically adjusts the size and architecture of the neural network to reflect the complexity and structure of the data. The compact network obtained by PrAda-net can be translated to additive model components, making it suitable for non-parametric statistical modelling with automatic model selection. We demonstrate PrAda-net on simulated data, where we compare the test error performance, variable importance and variable subset identification properties of PrAda-net to other lasso-based regularization approaches for neural networks. We also apply PrAda-net to the massive U.K. black smoke data set, to demonstrate how PrAda-net can be used to model complex and heterogeneous data with spatial and temporal components. In contrast to classical, statistical non-parametric approaches, PrAda-net requires no preliminary modeling to select the functional forms of the additive components, yet still results in an interpretable model representation.

Hybrid Statistical Models for Forecasting Yield of Mango and Banana in Tamil Nadu, India

Horticulture sector plays a prominent role in economic growth of India. India is the second largest producer of fruits and vegetables in the world next to China. Among the horticultural crops, fruit crops are cultivated in majority of the area in India. Fruit crops play a significant role in the economic development, nutritional security, employment generation, and total growth of country. India is major producer of mango and banana, among fruit crops. The objective of this research paper is to predicate the yield of mango and banana in Tamil Nadu using different models such as linear and nonlinear, parametric, and non-parametric statistical models. In this research, a hybrid model had been proposed, which consists of linear and nonlinear models. In this hybrid model, combination of the Autoregressive Integrated Moving Average (ARIMA) and Regression model were used. The present study was conducted in Tamil Nadu. Since, area and production of Mango and Banana are higher in Tamil Nadu. Based on results obtained production and yield of Mango and Banana were predicted for next four years.

Data-Driven Reinforcement-Learning-Based Automatic Bucket-Filling for Wheel Loaders

Automation of bucket-filling is of crucial significance to the fully automated systems for wheel loaders. Most previous works are based on a physical model, which cannot adapt to the changeable and complicated working environment. Thus, in this paper, a data-driven reinforcement-learning (RL)-based approach is proposed to achieve automatic bucket-filling. An automatic bucket-filling algorithm based on Q-learning is developed to enhance the adaptability of the autonomous scooping system. A nonlinear, non-parametric statistical model is also built to approximate the real working environment using the actual data obtained from tests. The statistical model is used for predicting the state of wheel loaders in the bucket-filling process. Then, the proposed algorithm is trained on the prediction model. Finally, the results of the training confirm that the proposed algorithm has good performance in adaptability, convergence, and fuel consumption in the absence of a physical model. The results also demonstrate the transfer learning capability of the proposed approach. The proposed method can be applied to different machine-pile environments.

Uncertainty analysis based on non-parametric statistical modelling method for photovoltaic array output and its application in fault diagnosis

Fault diagnosis is crucial for photovoltaic (PV) power generation, but the output uncertainty characteristics of PV arrays caused by different failures and their fluctuations make them facing new challenges. The parameter estimation (PE) method is widely used for uncertainty analysis, but there exist big differences between the PE results and the real output distributions. To solve this problem, this paper proposes a method for acquiring the fault diagnosis threshold based on non-parametric kernel density estimation (NKDE). First, the distribution characteristics of PV arrays output are statistically analysed, it is found that current and power are mainly affected by the solar irradiance, resulting in strong volatility and uncertainty, which makes it hard to get the threshold for fault diagnosis, thus we propose new three status indicators to finish it. Then, the probability models of the three indicators are built based on the kernel density estimation (KDE) method, by assigning the confidence values of the models, the fault diagnosis threshold intervals can be obtained. Verification shows that NKDE method performed better than traditional PE method in fitting the distribution of the PV arrays output, it does not need prior knowledge of the probability density function. And the proposed fault diagnosis method proves it is feasible to apply the uncertainty analysis method to PV array fault diagnosis. The paper provides a new idea for setting the dynamic threshold for PVarray fault diagnosis.

Spinal Muscular Atrophy after Nusinersen Therapy: Improved Physiology in Pediatric Patients with No Significant Change in Urine, Serum, and Liquor 1H-NMR Metabolomes in Comparison to an Age-Matched, Healthy Cohort.

Spinal muscular atrophy (SMA) is a genetically heterogeneous group of rare neuromuscular diseases and was until recently the most common genetic cause of death in children. The effects of 2-month nusinersen therapy on urine, serum, and liquor 1H-NMR metabolomes in SMA males and females were not explored yet, especially not in comparison to the urine 1H-NMR metabolomes of matching male and female cohorts. In this prospective, single-centered study, urine, serum, and liquor samples were collected from 25 male and female pediatric patients with SMA before and after 2 months of nusinersen therapy and urine samples from a matching healthy cohort (n = 125). Nusinersen intrathecal application was the first therapy for the treatment of SMA by the Food and Drug Administration (FDA) and the European Medicines Agency (EMA). Metabolomes were analyzed using targeted metabolomics utilizing 600 MHz 1H-NMR, parametric and nonparametric multivariate statistical analyses, machine learning, and modeling. Medical assessment before and after nusinersen therapy showed significant improvements of movement, posture, and strength according to various medical tests. No significant differences were found in metabolomes before and after nusinersen therapy in urine, serum, and liquor samples using an ensemble of statistical and machine learning approaches. In comparison to a healthy cohort, 1H-NMR metabolomes of SMA patients contained a reduced number and concentration of urine metabolites and differed significantly between males and females as well. Significantly larger data scatter was observed for SMA patients in comparison to matched healthy controls. Machine learning confirmed urinary creatinine as the most significant, distinguishing SMA patients from the healthy cohort. The positive effects of nusinersen therapy clearly preceded or took place devoid of significant rearrangements in the 1H-NMR metabolomic makeup of serum, urine, and liquor. Urine creatinine was successful at distinguishing SMA patients from the matched healthy cohort, which is a simple systemic novelty linking creatinine and SMA to the physiology of inactivity and diabetes, and it facilitates the monitoring of SMA disease in pediatric patients through non-invasive urine collection.

Improving the Performance of Index Insurance Using Crop Models and Phenological Monitoring

Extreme weather events cause considerable damage to the livelihoods of smallholder farmers globally. Whilst index insurance can help farmers cope with the financial consequences of extreme weather, a major challenge for index insurance is basis risk, where insurance payouts correlate poorly with actual crop losses. We analyse to what extent the use of crop simulation models and crop phenology monitoring can reduce basis risk in index insurance. Using a biophysical process-based crop model (Agricultural Production System sIMulator (APSIM)) applied for rice producers in Odisha, India, we simulate a synthetic yield dataset to train non-parametric statistical models to predict rice yields as a function of meteorological and phenological conditions. We find that the performance of statistical yield models depends on whether meteorological or phenological conditions are used as predictors and whether one aggregates these predictors by season or crop growth stage. Validating the preferred statistical model with observed yield data, we find that the model explains around 54% of the variance in rice yields at the village cluster (Gram Panchayat) level, outperforming vegetation index-based models that were trained directly on the observed yield data. Our methods and findings can guide efforts to design smart phenology-based index insurance and target yield monitoring resources in smallholder farming environments.

Randomization-based inference in the presence of selection bias.

For the analysis of clinical trials, the study participants are usually assumed to be representative sample of a target population. This assumption is rarely fulfilled in clinical trials, and particularly not if the sample size is small. In addition, covariate imbalances may affect the trial. Randomization tests provide a nonparametric analysis method of the treatment effect that does not rely on population-based assumptions. We propose a nonparametric statistical model that yields a formal basis for randomization tests. We adapt the model for the presence of covariate imbalance in the form of selection bias and investigate the effects of bias on the rejection probability of the randomization test using Monte Carlo simulations. Finally, we show that ancillary statistics can be used to control for the influence of bias. We show that covariate imbalance leads to an inflation of the type I error probability. The proposed nonparametric model allows for the use of ancillary statistics that yield an unbiased adjusted randomization test.