Statistical Learning Research Articles

Classifying PDO Kalamata Olive Oil from Geographic Origins of the Messenia Region based on Statistical Machine Learning

Kalamata is a smart city located in southeastern Greece in the Mediterranean basin and it is the capital of the Messenia regional unit. It is known for the famous Protected Designation of Origin (PDO) Kalamata olive oil produced mainly from the Koroneiki olive variety. The PDO Kalamata olive oil, established by Council regulation (EC) No 510/2006, owes its quality and special characteristics to the geographical environment, olive tree variety, and human factor. The PDO Kalamata olive oil is produced exclusively in the regional unit of Messenia, being the main profit of local farmers. However, soil chemical composition, microclimates, and agronomic factors are changed within the Messenia spatial area leading to differentiation of PDO Kalamata olive oil characteristic. In this paper, we use statistical machine learning algorithms to determine the geographical origin of Kalamata olive oil at PDO level based on synchronous excitation−emission fluorescence spectroscopy of olive oils. Evaluations of the statistical models are promising for differentiating the origin of PDO Kalamata olive oil with high values of prediction accuracy thus enabling companies that process and bottle kalamata olive oil to choose olive oil from a specific region of Messenia that fulfills certain characteristics. Concretely, the current research effort focuses on a specific olive oil variety within a limited geographic region. Intuitively, future research should also focus on validation of the proposed methodology to other olive oil varieties and production areas.

Effect of corrosion pit on fatigue damage and failure in powder bed fusion AlSi10Mg

AbstractThis paper is concerned with environmental corrosion during service of load‐bearing aluminum alloys fabricated by additive manufacturing (AM) with a focus on damage evolution characterization and identification of dominated failure mechanisms of pre‐corroded selective laser melted aluminum alloy. An experimental strategy for analysis of the damage and failure process is presented, combining 3D surface measurement, 3D digital image correlation, and scanning electron microscopy, to provide a multi‐source experimental characterization of corrosion morphology, strain field evolution, and fracture morphology. Statistics and machine learning methods were employed to process the measured multi‐source experimental data, showing that local average roughness has a strong influence on macro‐crack initiation position. The analysis focuses on four primary types of fatigue micro‐crack initiation, namely, internal defect, diffuse corrosion, corrosion micro‐pit, and corrosion jut; internal defect and diffuse corrosion are unique features for powder bed fusion (PBF) AlSi10Mg. The micro‐crack initiation mechanisms for all four types are (i) local stress concentrations due to adjacent material micro‐structural defects, (ii) interaction of underlying alloy microstructure and corrosion‐induced stress/strain, and (iii) local stress/strain concentration at corrosion micro‐pit and jut, respectively.

Entropy pelican optimization algorithm (epoa) based feature selection and deep autoencoder (dae) of heart failure status prediction

Introduction: Heart Failure (HF) is a complicated condition as well as a significant public health issue. Data processing is now required for machine and statistical learning techniques while it helps to identify key features and eliminates unimportant, redundant, or noisy characteristics, hence minimizing the feature space's dimensions. A common cause of mortality in cases of heart disease is Dilated Cardiomyopathy (DCM). Methods: The feature selection in this work depends on the Entropy Pelican Optimization Algorithm (EPOA). It is a recreation of pelicans' typical hunting behaviour. This is comparable to certain characteristics that lead to better approaches for solving high-dimensional datasets. Then Deep Autoencoder (DAE) classifier has been introduced for the prediction of patients. DAE classifier is employed to compute the system's nonlinear function through data from the normal and failure state. Results: DAE was discovered to not only considerably increase accuracy but also to be beneficial when there is a limited amount of labelled data.Performance metrics like recall, precision, accuracy, f-measure, and error rate has been used for results analysis. Conclusion: Publicly available benchmark dataset has been collected from Gene Expression Omnibus (GEO) repository to evaluate and contrast the suitability of the suggested classifier with other existing methods.

Using an mHealth approach to collect patient-generated health data for predicting adverse health outcomes among adult survivors of childhood cancer.

Cancer therapies predispose childhood cancer survivors to various treatment-related late effects, which contribute to a higher symptom burden, chronic health conditions (CHCs), and premature mortality. Regular monitoring of symptoms between clinic visits is useful for timely medical consultation and interventions that can improve quality of life (QOL). The Health Share Study aims to utilize mHealth to collect patient-generated health data (PGHD; daily symptoms, momentary physical health status) and develop survivor-specific risk prediction scores for mitigating adverse health outcomes including poor QOL and emergency room admissions. These personalized risk scores will be integrated into the hospital-based electronic health record (EHR) system to facilitate clinician communications with survivors for timely management of late effects. This prospective study will recruit 600 adult survivors of childhood cancer from the St. Jude Lifetime Cohort study. Data collection include 20 daily symptoms via a smartphone, objective physical health data (physical activity intensity, sleep performance, and biometric data including resting heart rate, heart rate variability, oxygen saturation, and physical stress) via a wearable activity monitor, patient-reported outcomes (poor QOL, unplanned healthcare utilization) via a smartphone, and clinically ascertained outcomes (physical performance deficits, onset of/worsening CHCs) assessed in the survivorship clinic. Participants will complete health surveys and physical/functional assessments in the clinic at baseline, 2) report daily symptoms, wear an activity monitor, measure blood pressure at home over 4 months, and 3) complete health surveys and physical/functional assessments in the clinic 1 and 2 years from the baseline. Socio-demographic and clinical data abstracted from the EHR will be included in the analysis. We will invite 20 cancer survivors to investigate suitable formats to display predicted risk information on a dashboard and 10 clinicians to suggest evidence-based risk management strategies for adverse health outcomes. Machine and statistical learning will be used in prediction modeling. Both approaches can handle a large number of predictors, including longitudinal patterns of daily symptoms/other PGHD, along with cancer treatments and socio-demographics. The individualized risk prediction scores and added communications between providers and survivors have the potential to improve survivorship care and outcomes by identifying early clinical presentations of adverse events.

Equivalent sentiment measures for cross-language analysis of corporate communications

This paper presents a technique for sentiment measurement in many languages. The method allows researchers to efficiently analyze corporate documents, management reports, and financial statements using python. When the texts are written in many languages, the method extracts equivalent cross-linguistic sentiment features that can be used for statistical analysis or machine learning. We use Open Multilingual WordNet, a large lexicon organizing words into semantic groups, as the knowledge base about word equivalence in more than 200 languages. We experiment with a parallel English-French corpus and find that our senitment measures across the two languages are comparable. The method produces a consistent classification of positive and negative texts in two languages, and sentiment measure values correlate. The paper provides a detailed account of the method and python code, So that it can be applied to other languages, text mining, quantitative communication studies, and management research.•Method to create equivalent sentiment measures in multiple languages•Based on established lexicons and WordNet•Validated for English and French

Semantic and Syntactic Predictions in Reading Aloud: Are Good Predictors Good Statistical Learners?

Recent research suggests that becoming a fluent reader may partially rely on a domain-general statistical learning (SL) mechanism that allows a person to automatically extract predictable patterns from the sensory input. The goal of the present study was to investigate a potential link between SL and the ability to make linguistic predictions. All previous studies investigated quite general levels of reading ability rather than the dynamic process of making linguistic predictions. We thus used a recently developed predictive reading task, which consisted of having participants read aloud words that were preceded by either semantically or syntactically predictive contexts. To measure the componential nature of SL, we used a visual and an auditory SL task (VSL, ASL) and the classic serial reaction time task (SRT). General reading ability was assessed with a reading speed/comprehension test. The study was conducted online on a sample of 120 participants to make it possible to explore interindividual differences. The results showed only weak and sometimes even negative correlations between the various SL measures. ASL correlated positively and predicted general reading ability but neither semantic nor syntactic prediction effects. Similarly, one of the SRT measures was significantly associated with reading level and reading speed but not with linguistic prediction effects. In sum, there is little evidence that domain-general SL is a good predictor of people's ability to make domain-specific linguistic predictions. In contrast, SL shows a weak but significant association with general reading ability.

A feature selection method based on Shapley values robust for concept shift in regression

AbstractFeature selection is one of the most relevant processes in any methodology for creating a statistical learning model. Usually, existing algorithms establish some criterion to select the most influential variables, discarding those that do not contribute to the model with any relevant information. This methodology makes sense in a static situation where the joint distribution of the data does not vary over time. However, when dealing with real data, it is common to encounter the problem of the dataset shift and, specifically, changes in the relationships between variables (concept shift). In this case, the influence of a variable cannot be the only indicator of its quality as a regressor of the model, since the relationship learned in the training phase may not correspond to the current situation. In tackling this problem, our approach establishes a direct relationship between the Shapley values and prediction errors, operating at a more local level to effectively detect the individual biases introduced by each variable. The proposed methodology is evaluated through various examples, including synthetic scenarios mimicking sudden and incremental shift situations, as well as two real-world cases characterized by concept shifts. Additionally, we perform three analyses of standard situations to assess the algorithm’s robustness in the absence of shifts. The results demonstrate that our proposed algorithm significantly outperforms state-of-the-art feature selection methods in concept shift scenarios, while matching the performance of existing methodologies in static situations.

Transfer of statistical learning between tasks.

Recent studies have shown that observers can learn to suppress locations in the visual field with a high distractor probability. Here, we investigated whether this learned suppression resulting from a spatial distractor imbalance transfers to a completely different search task that does not contain any distractors. Observers performed the additional singleton task and learned to suppress the location that was likely to contain a color singleton distractor. Within a block, the additional singleton task would randomly switch to a T-among-L task where observers searched in parallel (Experiment 1) or serially (Experiment 2) for a T among Ls. The upcoming search was either unpredictable (Experiment 1/2A) or cued (Experiment 1/2B). The results show that there was transfer of learning from one to the other task as the learned suppression stayed in place after the switch regardless of whether the T-among-L task was performed via parallel or serial search. Moreover, cueing that the task would switch had no effect on performance. The current findings indicate that implicit learned biases are rather inflexible and remain in place even when the task and the required search strategy are dramatically different and even when participants can anticipate that a change in the search required is imminent. This transfer of the suppression to a different task is consistent with the notion that suppression is proactively applied. Because the location is already suppressed proactively, that is, before display onset, regardless which display and task is presented, the suppressed location competes less for attention than all other locations. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Deducing neighborhoods of classes from a fitted model

AbstractIn this article, a new kind of interpretable machine learning method is presented, which can help to understand the partition of the feature space into predicted classes in a classification model using quantile shifts, and this way make the underlying statistical or machine learning model more trustworthy. Basically, real data points (or specific points of interest) are used and the changes of the prediction after slightly raising or decreasing specific features are observed. By comparing the predictions before and after the shifts, under certain conditions the observed changes in the predictions can be interpreted as neighborhoods of the classes with regard to the shifted features. Chord diagrams are used to visualize the observed changes. For illustration, this quantile shift method (QSM) is applied to an artificial example with medical labels and a real data example.

Selective Neural Entrainment Reveals Hierarchical Tuning to Linguistic Regularities in Reading

ABSTRACT Reading is both a visual and a linguistic task, and as such it relies on both general-purpose, visual mechanisms and more abstract, meaning-oriented processes. Disentangling the roles of these resources is of paramount importance in reading research. The present study capitalizes on the coupling of fast periodic visual stimulation and MEG recordings to address this issue and investigate the role of different kinds of visual and linguistic units in the visual word identification system. We compared strings of pseudo-characters; strings of consonants (e.g., sfcl); readable, but unattested strings (e.g., amsi); frequent, but non-meaningful chunks (e.g., idge); suffixes (e.g., ment); and words (e.g., vibe); and looked for discrimination responses with a particular focus on the ventral, occipito-temporal regions. The results revealed sensitivity to alphabetic, readable, familiar, and lexical stimuli. Interestingly, there was no discrimination between suffixes and equally frequent, but meaningless endings, thus highlighting a lack of sensitivity to semantics. Taken together, the data suggest that the visual word identification system, at least in its early processing stages, is particularly tuned to form-based regularities, most likely reflecting its reliance on general-purpose, statistical learning mechanisms that are a core feature of the visual system as implemented in the ventral stream.

Acute Stress Effects on Statistical Learning and Episodic Memory.

Stress is widely considered to negatively impact hippocampal function, thus impairing episodic memory. However, the hippocampus is not merely the seat of episodic memory. Rather, it also (via distinct circuitry) supports statistical learning. On the basis of rodent work suggesting that stress may impair the hippocampal pathway involved in episodic memory while sparing or enhancing the pathway involved in statistical learning, we developed a behavioral experiment to investigate the effects of acute stress on both episodic memory and statistical learning in humans. Participants were randomly assigned to one of three conditions: stress (socially evaluated cold pressor) immediately before learning, stress ∼15 min before learning, or no stress. In the learning task, participants viewed a series of trial-unique scenes (allowing for episodic encoding of each image) in which certain scene categories reliably followed one another (allowing for statistical learning of associations between paired categories). Memory was assessed 24 hr later to isolate stress effects on encoding/learning rather than retrieval. We found modest support for our hypothesis that acute stress can amplify statistical learning: Only participants stressed ∼15 min in advance exhibited reliable evidence of learning across multiple measures. Furthermore, stress-induced cortisol levels predicted statistical learning retention 24 hr later. In contrast, episodic memory did not differ by stress condition, although we did find preliminary evidence that acute stress promoted memory for statistically predictable information and attenuated competition between statistical and episodic encoding. Together, these findings provide initial insights into how stress may differentially modulate learning processes within the hippocampus.

Predictive Modelling of Postpartum Haemorrhage Using Early Risk Factors: A Comparative Analysis of Statistical and Machine Learning Models.

Postpartum haemorrhage (PPH) is a significant cause of maternal morbidity and mortality worldwide, particularly in low-resource settings. This study aimed to develop a predictive model for PPH using early risk factors and rank their importance in terms of predictive ability. The dataset was obtained from an observational case-control study in northern Rwanda. Various statistical models and machine learning techniques were evaluated, including logistic regression, logistic regression with elastic-net regularisation, Random Forests, Extremely Randomised Trees, and gradient-boosted trees with XGBoost. The Random Forest model, with an average sensitivity of 80.7%, specificity of 71.3%, and a misclassification rate of 12.19%, outperformed the other models, demonstrating its potential as a reliable tool for predicting PPH. The important predictors identified in this study were haemoglobin level during labour and maternal age. However, there were differences in PPH risk factor importance in different data partitions, highlighting the need for further investigation. These findings contribute to understanding PPH risk factors, highlight the importance of considering different data partitions and implementing cross-validation in predictive modelling, and emphasise the value of identifying the appropriate prediction model for the application. Effective PPH prediction models are essential for improving maternal health outcomes on a global scale. This study provides valuable insights for healthcare providers to develop predictive models for PPH to identify high-risk women and implement targeted interventions.

Energy Load Forecasting Techniques in Smart Grids: A Cross-Country Comparative Analysis

Energy management systems allow the Smart Grids industry to track, improve, and regulate energy use. Particularly, demand-side management is regarded as a crucial component of the entire Smart Grids system. Therefore, by aligning utility offers with customer demand, anticipating future energy demands is essential for regulating consumption. An updated examination of several forecasting techniques for projecting energy short-term load forecasts is provided in this article. Each class of algorithms, including statistical techniques, Machine Learning, Deep Learning, and hybrid combinations, are comparatively evaluated and critically analyzed, based on three real consumption datasets from Spain, Germany, and the United States of America. To increase the size of tiny training datasets, this paper also proposes a data augmentation technique based on Generative Adversarial Networks. The results show that the Deep Learning-hybrid model is more accurate than traditional statistical methods and basic Machine Learning procedures. In the same direction, it is demonstrated that more comprehensive datasets assisted by complementary data, such as energy generation and weather, may significantly boost the accuracy of the models. Additionally, it is also demonstrated that Generative Adversarial Networks-based data augmentation may greatly improve algorithm accuracy.

Using an ensemble modeling to predict the potential distribution and habitat suitability of caracal (Caracal caracal) in southwestern Iran

Anthropogenic activities and climate change are degrading wildlife habitats, making it crucial to assess suitability of remaining habitat for effective conservation and management of threatened wildlife. This study aimed to determine the variables influencing habitat selection and spatial distribution of caracal (Caracal caracal) in southwestern Iran from 2019 to 2020. Ensembles of Small Models (ESM) were used based on eight species distribution models that included a combination of statistical and machine learning methods (GLM, CTA, FDA, GBM, ANN, MARS, RF, and MaxEnt). The outputs of two ensemble models were hybridized by simple averaging. In ESM 1, distance to water resources was used as a representative of environmental variables and in ESM 2, annual precipitation was used as a representative of bioclimatic variables. In both ESMs, these variables were combined with ecological variables such as the distribution of wild goat (Capra aegagrus) and livestock. The average importance of the model variables for ESM 1 and ESM 2 showed that the distribution of wild goat and livestock were the most important factors in predicting suitable habitat for caracals, respectively. Habitat suitability was higher close to the distribution of wild goat and livestock, near water sources and during periods of higher precipitation. Probabilistic predictions of caracal presence based on the hybridization of ESM 1 and ESM 2 suggest that only 1.3 % (206 ha) of the study area was highly suitable habitat, while 58.4 % (9570 ha) was unsuitable. Based on the ensemble of binary presence-absence maps, caracals were certainly absent in 76.2 % (12,490 ha) of the area, while their presence was uncertain in 21.6 % (3536 ha). The presence of caracals was certain in 2.2 % (363 ha) of the study area. Awareness of the habitat and distribution of caracal can help to make better decisions towards conservation and management and also reduce its conflict with humans.

Knowledge-based system and expectation-maximization to discovering causes of imperfect labels in vehicular networks clustering

This paper focuses on the development of a knowledge-based system for automatically diagnosing issues in Vehicular ad hoc networks (VANETs). VANETs enable communication between vehicles and infrastructure, enhancing road safety and efficiency through timely information exchange. The proposed system aims to efficiently maintain and ensure the continuity of network service by leveraging innovative pattern recognition methods tailored to VANETs. The automatic diagnosis problem in VANETs involves estimating the operating class of network components based on sensor observations. This entails associating sensor measurements with specific operating modes. By implementing condition-based preventive maintenance procedures, potential component failures can be detected early, mitigating network disruptions. Various approaches, such as expert systems, fault trees, network state models, and statistical learning through pattern recognition, can be employed to address this problem. This paper primarily focuses on the statistical learning approach, where a classification or regression function is learned from a set of examples to assign operation modes to new measurements. It discusses relevant metrics and preprocessing techniques to simplify the decision-making process. The diagnostic system's results are determined based on the formulation of the classification or regression problem. The learning base is constructed, and an appropriate classification method is selected to develop and validate the automatic diagnosis system. While non-parametric models like support vector machines are commonly used, this article emphasizes the significance of considering assumptions and leveraging additional information to enhance performance. It proposes a more specific formalization of the problem, integrating the unique characteristics of VANETs. The contributions of this article revolve around the theory of belief functions, a generative approach, and the utilization of parametric models defined using graphical models. Experimental studies conducted on artificial datasets have demonstrated the benefits of the semi-supervised approach within the context of VANET networks.

A study on employment sustainability among Engineering students using a Statistical and Deep Learning framework

A study on employment sustainability among Engineering students using a Statistical and Deep Learning framework

An Enhanced Long-Term Wind Speed Prediction Using Dynamic Unified Ensemble Learning and Data Assimilation Techniques: A Case Study in Tamil Nadu, India

Reliable wind speed prediction is essential for effective grid management since wind energy is a key component of renewable energy generation. However, controlling wind speed just right is no easy feat since the wind flow changes constantly. This paper presents an innovative method for predicting wind speeds in the future. It uses state-of-the-art data assimilation methods and a dynamic unified ensemble learning model. For efficient wind energy planning and monitoring, WSP accuracy is crucial. Data from a single site limited the accuracy of WSPs in previous research. An improved accuracy of long-term wind speed predictions is achieved by the proposed model via the integration of data assimilation and ensemble learning. To increase forecast accuracy, the model uses sophisticated data assimilation methods like the Kalman filter to combine data from many sources. Specifically, the model employs the Stacked CNN + BiLSTM with Data Assimilator (SCBLSTM+DA) technique, which integrates Wind Speed (WS) data from adjacent areas with the CNN + BiLSTM-based Ensemble Learning Model (ELM) and the Four-Dimensional Variational and Ensemble Kalman Filter (4DVar/EnKF) Data Assimilation method. Using real-world wind speed data from nine meteorological stations in Tamil Nadu, India, we find that current prediction models, including both classical statistical and cutting-edge machine learning models, perform better. Further, unlike standalone models, the suggested model shows less susceptibility to changes in prediction time scales. Promising a solution to improve long-term wind speed predictions accuracy, this study has significant consequences for wind energy management and production.

Employing Graph Neural Networks for Predicting Electrode Average Voltages and Screening High-Voltage Sodium Cathode Materials.

For many years, humans have been relentlessly focused on enhancing battery longevity and boosting energy storage capacities. The performance and durability of a battery depend significantly on the material used for its electrodes. In this context, merging machine learning with density functional theory (DFT) calculations has emerged as a pivotal approach to advancing the exploration of battery crystal structures. We present a new method that combines a graph convolutional neural network (GNN) with a Transformer convolutional layer, which we call Transformer-GNN. To underscore its efficacy, we benchmarked Transformer-GNN against three established statistical machine learning models: Support Vector Machine, Random Forest, and XGBoost. We also developed a standard GNN, which we refer to as Basic-GNN. Additionally, we compared Basic-GNN with Transformer-GNN to highlight the improvements brought about by incorporating the Transformer convolutional layer. The Transformer-GNN model outperforms the other models, achieving the highest R2 value of 0.82 and the lowest mean squared error of 0.3161. Our findings demonstrate that the Transformer-GNN can profoundly understand battery crystal structures, thus forging the path toward more sophisticated and durable battery systems. Leveraging the GNN model's voltage predictions in tandem with the capacity data sourced from the database, we screened and pinpointed Na(NiO2)2 as a high-voltage (higher than 5 V), high-capacity sodium cathode material. We conducted DFT calculations on Na(NiO2)2 and revealed the migration mechanism of the Na ions.

Multiple Data Sources in Reliability Assessment of Distribution Networks

Since distribution networks provide the regular and secure transmission of power to consumers, its dependability is essential to the smooth operation of modern society. Conventional techniques for evaluating these networks' dependability have depended on a single data source, such as historical outage statistics. The capacity to combine many data sources into dependability evaluation has been possible due to technological improvements and the abundance of available data. This allows for a more thorough and precise understanding of network performance. This technical abstract proposes a methodology for incorporating multiple data sources in the reliability assessment of distribution networks. It is to identify and collect various data sources such as weather data, customer complaints, equipment maintenance records, and outage data. These data sources are then integrated and analyzed using statistical and machine learning algorithms to identify patterns and correlations that can provide insights into the network's reliability.

Statistical learning and mathematics knowledge: the case of arithmetic principles

Statistical learning—an unconscious cognitive process used to extract regularities—is well-established as a fundamental mechanism underlying learning. Yet, despite the prominence of patterns in the number system and operations, little is known about the relation between statistical learning and mathematics knowledge. This study examined the associations among statistical learning, executive control, and arithmetic knowledge among first graders (N = 54). The relations varied by operation. For addition, children with greater statistical learning capacity responded more quickly to problems that were part of a principle (i.e., commutativity) pair than to unrelated problems, even after accounting for baseline performance, executive control, and age. For subtraction, results indicated an interaction between children's baseline subtraction performance and their statistical learning on accuracy. These findings provide an impetus for testing new models of mathematics learning that include statistical learning as a potentially important mechanism.