Self-organizing Map Research Articles

A new driving style recognition method for personalized adaptive cruise control to enhance vehicle personalization

There are limitations of personalization in Advanced Driver Assistance Systems (ADAS) that have a serious impact on driver acceptance and satisfaction. This study investigates driving style recognition method to achieve personalization of longitudinal driving behavior. Currently, driving style recognition algorithms for Personalized Adaptive Cruise Control (PACC) rely on integrated recognition. However, disturbances in the driving cycle may lead to changes in a driver’s integrated driving style. Therefore, the integrated driving style cannot accurately and comprehensively reflect the driver’s driving style. To solve this problem, a new driving style recognition method for PACC is proposed, which considers integrated driving style and driving cycle. Firstly, the method calculates the constructed feature parameters of driving cycle and style, and then reduces the dimensionality of the feature parameter matrix by principal component analysis (PCA). Secondly, a two-stage clustering algorithm with self-organizing mapping networks and K-means clustering (SOM-K-means) is used to obtain the type labels. Then, a transient recognition model based on random forest (RF) is established and the hyperparameters of this model are optimized by sparrow search algorithm (SSA). Based on this, a comprehensive driving style recognition model is established using analytic hierarchy process (AHP). Finally, the validity of the proposed method is verified by a natural dataset. The method incorporates the driving cycle into driving style recognition and provides guidance for improving the personalization of adaptive cruise control system.

Joint effects of indoor air pollution and maternal psychosocial factors during pregnancy on trajectories of early childhood psychopathology.

Prenatal indoor air pollution and maternal psychosocial factors have been associated with adverse psychopathology. We used environmental-exposure mixture methodology to investigate joint effects of both exposure classes on child behavior trajectories. For 360 children from the South African Drakenstein Child Health Study, we created trajectories of Child Behavior Checklist scores (at 24, 42, and 60 months) using latent-class linear mixed effects models. Indoor air pollutants and psychosocial factors were measured during pregnancy (second trimester). After adjusting for confounding, single-exposure effects (per natural log-1 unit increase) were assessed using polytomous logistic regression models, joint effects using self-organizing maps, and principal component analysis. Three trajectories were chosen for both internalizing and externalizing problems, with "high" (externalizing) or "increasing" (internalizing) being the most adverse trajectories. High externalizing trajectory was associated with increased exposure to particulate matter of ≤ 10 microns in diameter (PM10) (odds ratio [OR] = 1.25; 95% CI, 1.01-1.55) and self-organizing maps exposure profile most associated with smoking (OR = 2.67; 95% CI, 1.14-6.27). Medium internalizing trajectory was associated with increased emotional intimate partner violence (OR = 2.66; 95% CI, 1.17-5.57), increasing trajectory with increased benzene (OR = 1.24; 95% CI, 1.02-1.51) and toluene (1.21; 95% CI, 1.02-1.44) and the principal component most correlated with benzene and toluene (OR = 1.25; 95% CI, 1.02-1.54). Prenatal exposure to environmental pollutants and psychosocial factors was associated with internalizing and externalizing child behavior trajectories. Understanding joint effects of adverse exposure mixtures will facilitate targeted interventions to prevent childhood psychopathology. This article is part of a Special Collection on Mental Health.

Unveiling the evolution of urban rail transit network: considering ridership attributes

ABSTRACT Previous research on urban rail transit (URT) evolution mainly focused on network topology, neglecting ridership attributes. This study extracts ridership and network topology indicators from Chinese URT data. Employing a self-organizing mapping neural network model, it divides China’s URT development into four stages. The initial stage and the development stage form the framework of URT network. The network diameter reaches the maximum in the networked operation stage. In the mature stage, URT network densification occurs alongside a significant increase in resident ridership. It is also found that each network indicator has a significant nonlinear relationship with ridership attributes. These findings are of guiding significance for urban planners to accurately understanding URT’s future development and rational network planning and construction.

Self-organizing maps applied to the analysis and identification of characteristics related to air quality monitoring stations and its pollutants

Self-organizing maps applied to the analysis and identification of characteristics related to air quality monitoring stations and its pollutants

Validating the knowledge represented by a self-organizing map with an expert-derived knowledge structure.

Professionals are reluctant to make use of machine learning results for tasks like curriculum development if they do not understand how the results were generated and what they mean. Visualizations of peer reviewed medical literature can summarize enormous amounts of information but are difficult to interpret. This article reports the validation of the meaning of a self-organizing map derived from the Medline/PubMed index of peer reviewed medical literature by its capacity to coherently summarize the references of a core psychiatric textbook. Reference lists from ten editions of Kaplan and Sadock's Comprehensive Textbook of Psychiatry were projected onto a self-organizing map trained on Medical Subject Headings annotating the complete set of peer reviewed medical research articles indexed in the Medline/PubMed database (MedSOM). K-means clustering was applied to references from every edition to examine the ability of the self-organizing map to coherently summarize the knowledge contained within the textbook. MedSOM coherently clustered references into six psychiatric knowledge domains across teneditions (1967-2017). Clustering occurred at the abstract level of broad psychiatric practice including General/adult psychiatry, Child psychiatry, and Administrative psychiatry. The uptake of visualizations of published medical literature by medical experts for purposes like curriculum development depends upon validation of the meaning of the visualizations. The current research demonstrates that a self-organizing map (MedSOM) can validate the stability and coherence of the references used to support the knowledge claims of a standard psychiatric textbook, linking the products of machine learning to a widely accepted standard of knowledge.

A Multi-Layered Defence Strategy against DDoS Attacks in SDN/NFV-Based 5G Mobile Networks

Software-defined networking (SDN) and network functions virtualisation (NFV) are crucial technologies for integration in the fifth generation of cellular networks (5G). However, they also pose new security challenges, and a timely research subject is working on intrusion detection systems (IDSs) for 5G networks. Current IDSs suffer from several limitations, resulting in a waste of resources and some security threats. This work proposes a new three-layered solution that includes forwarding and data transport, management and control, and virtualisation layers, emphasising distributed controllers in the management and control layer. The proposed solution uses entropy detection to classify arriving packets as normal or suspicious and then forwards the suspicious packets to a centralised controller for further processing using a self-organising map (SOM). A dynamic OpenFlow switch relocation method is introduced based on deep reinforcement learning to address the unbalanced burden among controllers and the static allocation of OpenFlow switches. The proposed system is analysed using the Markov decision process, and a Double Deep Q-Network (DDQN) is used to train the system. The experimental results demonstrate the effectiveness of the proposed approach in mitigating DDoS attacks, efficiently balancing controller workloads, and reducing the duration of the balancing process in 5G networks.

Heavy metal pollution and ecological risk assessment: A study on Linli County soils based on self-organizing map and positive factorization approaches

Heavy metal pollution and ecological risk assessment: A study on Linli County soils based on self-organizing map and positive factorization approaches

The Quantitative Role of Moisture and Vertical Motion in Shaping Summer Heavy Rainfall over North China under Two Distinct Large-Scale Weather Patterns

Abstract Two distinct categories of weather patterns, denoted as Type 1 and Type 2, which show higher-than-expected frequency of summer heavy rainfall days (HRDs) over North China (NC), are selected from nine weather patterns categorized by the self-organizing map algorithm during 1979–2019. The respective HRDs over NC exhibit dissimilar characteristics, with Type 1 showing a northern distribution and Type 2 a southern distribution. The quantitative disparities in terms of moisture content and vertical motion are discussed in reactions to the synoptic-scale patterns associated with HRDs. The outcomes of a 20-day backward tracking, using the so-called Water Accounting Model-2layers, reveal noteworthy contrasts in moisture sources. Type 1 predominantly receives moisture from the western North Pacific, while Type 2 relies more on contributions from the Arabian Sea, Bay of Bengal, and Eurasia. However, the major moisture sources with grid cells contributing more than 0.01 mm show a consistent cumulative contribution of 77% for Type 1 and 80% for Type 2. The finding suggests that the discrepancy between the two types cannot be solely attributed to moisture supply. Further examination of the transverse and shearwise Q-vector components provides insights into how these distinct weather patterns influence HRDs by the alteration of vertical motion. In Type 1, an upper-level jet entrance induces a thermally direct secondary circulation that enhances vertical motion, while a baroclinic trough plays a dominant role in generating vertical motion in Type 2. Moreover, these unique configurations for each type of weather pattern are not only pre-existing but also intensified during HRDs.

Unraveling the links between development, growth, and tourism specialization. A country-panel analysis using self-organizing maps

This research explores the relationship between development, economic growth and tourism specialisation in 123 countries between 1995 and 2019. Based on the hypothesis that a single model is inadequate to capture the complex dynamics between these variables, we classify countries before studying the link within each category. The empirical strategy is based on the application of an objective country classification procedure, known as Kohonen’s self-organising maps. This methodology, novel for this field of study, allows countries to be grouped based on the dynamics of several variables. Three distinct groups emerge from this process. We find that causality from tourism specialization to GDP growth and from tourism to human development are observed in low-income countries. While no reverse causality between HDI and tourism is observed in any of the specifications, causality from GDP growth to tourism is verified in the group of high-income countries with heightened tourism specialisation and development.

Efficiency of clustering methods and self-organizing maps of Adriatic sardines and anchovies regarding organochlorine and fatty acid burden.

The Adriatic Sea plays a crucial role as both a significant fishing ground and a thriving trading market for small pelagic edible fish. Recognized for their nutritional value, these fish are esteemed for their high protein content and abundance of polyunsaturated omega-3 and omega-6 fatty acids, making them a sought-after and healthful food choice. Nevertheless, pelagic species can also serve as a reservoir for lipophilic organochlorine pollutants, posing potential risks to human health. In this study, we compared traditional classification methods traditional principal component analysis (PCA) and Ward's clustering with an advanced self-organizing map (SOM) algorithm in determining distribution patterns of 24 organochlorines and 19 fatty acids in sardine and anchovy samples taken from the eastern Adriatic. The outcomes reveal the strengths and weaknesses of the three approaches (PCA, Ward's clustering, and SOM). However, it is evident that SOM has proven to be the most effective in offering detailed information and data visualization. Although sardines and anchovies exhibit similar distribution patterns for p,p'-DDE, PCB-28, PCB-138, PCB-153, PCB-118, and PCB-170, they differ in the concentrations of fatty acids such as stearic, palmitic, myristic, oleic, docosapentaenoic, and docosahexaenoic acid. Our findings supply valuable insights for environmental authorities and fish consumers concerning the potential risks associated with organochlorines in these two types of fish.

Predictive modeling of garlic quality in hybrid infrared-convective drying using artificial neural networks

Artificial Neural Networks (ANN) application for enhancing food drying processes has been gaining traction within the food industry, particularly because of its potential to optimize conditions while preserving quality. This study delves into the effects of varying convective temperatures (40, 50, 60 °C), air velocities (0.7, 1.0, and 1.5 m/s), and infrared radiation intensities (1500, 2000, 3000 W/m²) on the drying efficiency and quality of garlic slices. Additionally, it explores the capability of ANN to predict optimal drying conditions to balance time efficiency and quality retention. Experimental observations revealed that a combination of a 60°C air temperature, 1.0 m/s air velocity, and 3000 W/m² infrared radiation minimized the drying time to 4.5 h. However, this setting also resulted in a 13.4 % reduction in Allicin content and a decrease in flavor intensity to 4.09 mg/g, underscoring the complexity of optimizing drying conditions without compromising key quality attributes. The study highlights the intricate relationships between drying parameters and garlic quality through precise ANN modeling, which achieved an exceptional fit. Principal Component Analysis (PCA) further elucidated the inverse correlation between drying time and critical quality factors of color, Allicin content, and flavor. Utilizing the Self-Organizing Map (SOM) technique, the study identified five distinct optimization zones, suggesting that lower temperatures, intermediate infrared levels, and higher air velocities present an optimal balance for enhancing garlic slice quality while reducing drying time. This research underscores the potential of advanced computational tools in refining food drying processes, offering valuable insights for the food industry in its quest to improve efficiency without sacrificing product quality.

Measurement of aero-engine feature-hierarchy fusion degradation trend based on parameter-adaptive VMD method and improved transformer model

The accumulation of operational time in aero-engines leads to irreversible mechanical wear and tear, necessitating accurate measurement of the health evolution trend for effective predictive maintenance, thus reducing the risk of accidents and ensuring personnel safety. In this paper, a parameter-adaptive variational mode decomposition (VMD) method and improved transformer model are proposed to forecast the degradation trend of aero-engine feature hierarchy fusion. Firstly, in order to quantitatively evaluate the engine health evolution process, the health state aggregate indicator (HSAI) is innovatively constructed by employing the deep blend auto-encoder and self-organizing map network, which facilitate the feature-hierarchy fusion of multi-source sensory data. Secondly, for the significant characteristics with nonlinearity and stochastic fluctuation of the HSAI sequence, the multiscale frequency features are extracted by the parameter-adaptive VMD method with the improved gray wolf optimizer, which analyzes the inherent degradation law. Finally, considering the problem of parameter sharing in the transformer model, a simplified mixture of experts routing algorithm is introduced to implement the switch transformer model to further measure the future aero-engine health trends. Extensive experiments on the multi-source dataset of aero-engine confirm that the proposed method accomplishes the more superior performance for health evolution measurement compared with other available methods.

Dynamic impact of hybrid wind-solar photovoltaic power injection on small signal stability of Nigerian 11kV power system using Self Organizing Map neural network

Power industry reorganization, legislation on greenhouse gas emissions and inadequate conventional generation capacity encourage the idea of incorporating Distributed Generation (DG) into the distribution networks as unconventional sources of power generation. High penetration of DG has further increased the complexity of the analysis of small signal stability because the variability of renewable energy sources affects the stability of power system. Therefore, this research work investigates the effect of wind and solar photovoltaic power injection on small signal stability of Nigerian 11 kV power network. Wind and solar photovoltaic were modelled using probabilistic methods and output power at twenty-one different locations within Nigeria was computed. Using the real and reactive power data of the network, Self-Organizing Map (SOM) neural network was trained and tested for the assessment of the small signal stability of the distribution network before and after the injection of renewable power in python virtual environment. The integration of hybrid wind-solar PV distributed generation at 30 %, 60 %, 90 % and 100 % into the studied 11 kV power distribution network improved its small signal stability by 1.4 %, 2.1 %, 2.5 % and 3.2 % respectively. Moreover, increase in the injection of power gotten from wind and solar sources reduces the oscillation frequency of the system from 1.39 Hz to 1.01 Hz and increases the damping from 4.6 % to 13.2 %. Reactive power increment in the range of 5 %-20 % improved the small signal stability of the system by the range 2.4 %-7 %. This improvement will translate to better power transfer capability of the network.

Spatial response of water level and quality shows more significant heterogeneity during dry seasons in large river-connected lakes.

The spatial response mechanism of hydrology and water quality of large river-connected lakes is very complicated. In this study, we developed a spatial response analysis method that couples wavelet correlation analysis (WTC) with self-organizing maps (SOM), revealing the spatial response and variation of water level and water quality in Poyang Lake, China's largest river-connected lake, over the past decade. The results show that: (1) there was significant spatial heterogeneity in water level and quality during the dry seasons (2010-2018) compared to other hydrological stages. (2) We identified a more pronounced difference in response of water level and quality between northern and southern parts of Poyang Lake. As the distance increases from the northern lake outlet, the impact of rising water levels on water quality deterioration intensified during the dry seasons. (3) The complex spatial heterogeneity of water level and quality response in the dry seasons is primarily influenced by water level fluctuations from the northern region and the cumulative pollutant entering the lake from the south, which particularly leads to the reversal of the response in the central area of Poyang Lake. The results of this study can contribute to scientific decision-making regarding water environment zoning management in large river-connected lakes amidst complex environment conditions.

Chemical space deconstruction-based dynamic model ensemble architecture for molecular property prediction

Green solvents, catalysts, functional materials, drugs, and other chemical products now have a quick engine for design thanks to machine learning (ML) based prediction of molecular properties. However, the accuracy and stability of ML-based models can be impeded by poor data quality, which is rarely studied in chemical product discovery and design. Inspired by the dynamic ensemble selection (DES), an improved DES based on chemical space deconstruction is proposed in this work to accommodate the prediction task of molecular properties. We innovatively developed a chemical space representation and deconstruction model based on a self-organizing mapping (SOM) neural network, facilitating the rapid implementation of the improved DES on molecular samples. Consequently, a novel dynamic model ensemble architecture (SOM-DES) is proposed as a model enhancement technology to build a more accurate and stable ensemble model, aiming to improve the predictive performance on the chemical subspace within poor-quality data. To achieve the architecture, a supervised dimensionality reduction algorithm has been improved to enhance the deep mining of molecular feature information for DES optimization. Additionally, a novel resampling strategy based on the combination of the geometric synthetic minority oversampling technique (G-SMOTE) algorithm and chemical space deconstruction, as a data augmentation technology, has been proposed for mitigating the disadvantage of unbalanced data during DES training. The prediction task for enthalpy of formation of the ideal gas is applied as a case study to demonstrate the superiority of the proposed SOM-DES. The results indicate that the proposed SOM-DES (R2 = 0.9731, RMSE = 55.4639) outperforms the traditional static ensemble strategy (SES, R2 = 0.9552, RMSE = 71.5045) in terms of global chemical spatial precision. More importantly, for chemical subspaces that are difficult to predict due to low data quality, SOM-DES shows a significant reduction in prediction errors compared to SES.

Groundwater hydrochemical signatures, nitrate sources, and potential health risks in a typical karst catchment of North China using hydrochemistry and multiple stable isotopes.

Nitrate pollution in aquatic ecosystems has received growing concern, particularly in fragile karst basins. In this study, hydrochemical compositions, multiple stable isotopes (δ2H-H2O, δ18Ο-Η2Ο, δ15Ν-ΝΟ3-, and δ18Ο-ΝΟ3-), and Bayesian stable isotope mixing model (MixSIAR) were applied to elucidate nitrate pollution sources in groundwater of the Yangzhuang Basin. The Durov diagram identified the dominant groundwater chemical face as Ca-HCO3 type. The NO3- concentration ranged from 10.89 to 90.45mg/L (average 47.34mg/L), showing an increasing trend from the upstream forest and grassland to the downstream agricultural dominant area. It is worth noting that 47.2% of groundwater samples exceeded the NO3- threshold value of 50mg/L for drinking water recommended by the World Health Organization. The relationship between NO3-/Cl- and Cl- ratios suggested that most groundwater samples were located in nitrate mixed endmember from agricultural input, soil organic nitrogen, and manure & sewage. The Self-Organizing Map (SOM) and Pearson correlations analysis further indicated that the application of calcium fertilizer, sodium fertilizer, and livestock and poultry excrement in farmland elevated NO3- level in groundwater. The output results of the MixSIAR model showed that the primary sources of NO3- in groundwater were soil organic nitrogen (55.3%), followed by chemical fertilizers (28.5%), sewage & manure (12.7%), and atmospheric deposition (3.4%). Microbial nitrification was a dominant nitrogen conversion pathway elevating NO3- levels in groundwater, while the denitrification can be neglectable across the study area. The human health risk assessment (HHRA) model identified that about 88.9%, 77.8%, 72.2%, and 50.0% of groundwater samples posing nitrate's non-carcinogenic health hazards (HQ > 1) through oral intake for infants, children, females, and males, respectively. The findings of this study can offer useful biogeochemical information on nitrogen pollution in karst groundwater to support sustainable groundwater management in similar human-affected karst regions.

Unveiling the hub genes in the SIGLECs family in colon adenocarcinoma with machine learning.

Despite the recognized roles of Sialic acid-binding Ig-like lectins (SIGLECs) in endocytosis and immune regulation across cancers, their molecular intricacies in colon adenocarcinoma (COAD) are underexplored. Meanwhile, the complicated interactions between different SIGLECs are also crucial but open questions. We investigate the correlation between SIGLECs and various properties, including cancer status, prognosis, clinical features, functional enrichment, immune cell abundances, immune checkpoints, pathways, etc. To fully understand the behavior of multiple SIGLECs' co-evolution and subtract its leading effect, we additionally apply three unsupervised machine learning algorithms, namely, Principal Component Analysis (PCA), Self-Organizing Maps (SOM), K-means, and two supervised learning algorithms, Least Absolute Shrinkage and Selection Operator (LASSO) and neural network (NN). We find significantly lower expression levels in COAD samples, together with a systematic enhancement in the correlations between distinct SIGLECs. We demonstrate SIGLEC14 significantly affects the Overall Survival (OS) according to the Hazzard ratio, while using PCA further enhances the sensitivity to both OS and Disease Free Interval (DFI). We find any single SIGLEC is uncorrelated to the cancer stages, which can be significantly improved by using PCA. We further identify SIGLEC-1,15 and CD22 as hub genes in COAD through Differentially Expressed Genes (DEGs), which is consistent with our PCA-identified key components PC-1,2,5 considering both the correlation with cancer status and immune cell abundance. As an extension, we use SOM for the visualization of the SIGLECs and show the similarities and differences between COAD patients. SOM can also help us define subsamples according to the SIGLECs status, with corresponding changes in both immune cells and cancer T-stage, for instance. We conclude SIGLEC-1,15 and CD22 as the most promising hub genes in the SIGLECs family in treating COAD. PCA offers significant enhancement in the prognosis and clinical analyses, while using SOM further unveils the transition phases or potential subtypes of COAD.

Data dimensionality reduction for an optimal switching mode classification applied to a step-down power converter

Abstract A dimensional reduction algorithm is applied to an intelligent classification model with the purpose of improving the efficiency and accuracy. The proposed classification model, used to distinguish the operating mode: Hard- and Soft-Switching, is presented and an analysis of the synchronized rectified step-down converter is done. With the aim of improving the accuracy and reducing the computational cost of the model, three different methods for dimensional reduction are applied to the input dataset of the model: self-organizing maps, principal component analysis and correlation matrix. The obtained results show how the number of variable is highly reduced and the performance of the classification model is boosted: the results manifest an improve in the accuracy and efficiency of the classification.

Predicting UPDRS in Parkinson’s disease using ensembles of self-organizing map and neuro-fuzzy

Parkinson's Disease (PD) is a complex, degenerative disease that affects nerve cells that are responsible for body movement. Artificial Intelligence (AI) algorithms are widely used to diagnose and track the progression of this disease, which causes symptoms of Parkinson's disease in its early stages, by predicting the results of the Unified Parkinson's Disease Rating Scale (UPDRS). In this study, we aim to develop a method based on the integration of two methods, one complementary to the other, Ensembles of Self-Organizing Map and Neuro-Fuzzy, and an unsupervised learning algorithm. The proposed method relied on the higher effect of the variables resulting from the analysis of the initial readings to obtain a correct and accurate preliminary prediction. We evaluate the developed approach on a PD dataset including speech cues. The process was evaluated with root mean square error (RMSE) and modified R square (modified R2). Our findings reveal that the proposed method is effective in predicting UPDRS outcomes by a combination of speech signals (measures of hoarseness). As the preliminary results during the evaluation showed numbers that proved the worth of the proposed method, such as UPDRS = 0.955 and RMSE approximately 0.2769 during the prediction process.

The novel hierarchical clustering approach using self‐organizing map with optimum dimension selection

AbstractIntroductionData clustering is an important field of machine learning that has applicability in wide areas, like, business analysis, manufacturing, energy, healthcare, traveling, and logistics. A variety of clustering applications have already been developed. Data clustering approaches based on self‐organizing map (SOM) generally use the map dimensions (of the grid) ranging from 2 × 2 to 8 × 8 (4–64 neurons [microclusters]) without any explicit reason for using the particular dimension, and therefore optimized results are not obtained. These algorithms use some secondary approaches to map these microclusters into the lower dimension (actual number of clusters), like, 2, 3, or 4, as the case may be, based on the optimum number of clusters in the specific data set. The secondary approach, observed in most of the works, is not SOM and is an algorithm, like, cut tree or the other.MethodsIn this work, the proposed approach will give an idea of how to select the most optimal higher dimension of SOM for the given data set, and this dimension is again clustered into the lower actual dimension. Primary and secondary, both utilize the SOM to cluster the data and discover that the weight matrix of the SOM is very meaningful. The optimized two‐dimensional configuration of SOM is not the same for every data set, and this work also tries to discover this configuration.ResultsThe adjusted randomized index obtained on the Iris, Wine, Wisconsin diagnostic breast cancer, New Thyroid, Seeds, A1, Imbalance, Dermatology, Ecoli, and Ionosphere is, respectively, 0.7173, 0.9134, 0.7543, 0.8041, 0.7781, 0.8907, 0.8755, 0.7543, 0.5013, and 0.1728, which outperforms all other results available on the web and when no reduction of attributes is done in this work.ConclusionsIt is found that SOM is superior to or on par with other clustering approaches, like, k‐means or the other, and could be used successfully to cluster all types of data sets. Ten benchmark data sets from diverse domains like medical, biological, and chemical are tested in this work, including the synthetic data sets.