Use Of Self-organizing Maps Research Articles

Use of self-organizing maps for the classification of cardiometabolic risk and physical fitness in adolescents

This study aimed to automatically classify physical fitness and cardiometabolic risk in a Chilean adolescent using self-organizing maps. This cross-sectional study analysed a nationally representative database from the Physical Education Quality Measurement System (n = 7197). Physical fitness and cardiometabolic risk variables were derived from anthropometric indicators. Self-Organizing maps (SOM) were employed to identify participant profiles based on an unsupervised predictive model. After implementing and training the SOM, a detailed analysis of the generated maps was conducted to interpret the revealed relationships and clusters. The analysis resulted in three classification groups, categorizing the sample into low, moderate, and high-risk levels. Students with better physical fitness exhibited lower cardiometabolic risk levels and a lower body mass index. SOM, through an unsupervised model, is a reliable tool for classifying cardiometabolic risk and physical fitness in adolescents.

Estimating Galaxy Parameters with Self-organizing Maps and the Effect of Missing Data

The current and upcoming large data volume galaxy surveys require the use of machine-learning techniques to maximize their scientific return. This study explores the use of Self-Organizing Maps (SOMs) to estimate galaxy parameters with a focus on handling cases of missing data and providing realistic probability distribution functions for the parameters. We train an SOM with a simulated mass-limited lightcone assuming a ugrizY JHK s +IRAC data set, mimicking the Hyper Suprime-Cam Deep joint data set. For parameter estimation, we derive SOM likelihood surfaces considering photometric errors to derive total (statistical and systematic) uncertainties. We explore the effects of missing data, including which bands are particularly critical to the accuracy of the derived parameters. We demonstrate that the parameter recovery is significantly better when the missing bands are “filled in” rather than if they are completely omitted. We propose a practical method for such recovery of missing data.

Regionalization of the Onset and Offset of the Rainy Season in Senegal Using Kohonen Self-Organizing Maps

This study explores the spatiotemporal variability of the onset, end, and duration of the rainy season in Senegal. These phenological parameters, crucial for agricultural planning in West Africa, exhibit high interannual and spatial variability linked to precipitation. The objective is to detect and spatially classify these indices across Senegal using different approaches. Daily precipitation data and ERA5 reanalyses from 1981 to 2018 were utilized. The employed method enables the detection of key dates. Subsequently, the Kohonen algorithm spatially classifies these indices on topological maps. The results indicate a meridional gradient of the onset, progressively later from the southeast to the northwest, whereas the end follows a north–south gradient. The duration varies from 45 days in the north to 150 days in the south. The use of self-organizing maps allows for classifying the onset, end, and duration of the season into four zones for the onset and end, and three zones for the duration of the season. They highlight the interannual irregularity of transitions, with both early and late years. The dynamic analysis underscores the complex influence of atmospheric circulation fields, notably emphasizing the importance of low-level monsoon flux. These findings have tangible implications for improving seasonal forecasts and agricultural activity planning in Senegal. They provide information on the onset, end, and duration classes for each specific zone, which can be valuable for planning crops adapted to each region.

Seasonal and El Niño–Southern Oscillation-related ocean variability in the Panama Bight

Abstract. In the Panama Bight, two different seasonal surface circulation patterns coincide with a strong mean sea level variation, as observed from 27 years of absolute dynamic topography (ADT) and the use of self-organizing maps. From January to April, a cyclonic gyre with a strong southwestward Panama Jet Surface Current (PJSC) dominates the basin circulation, forced by the Panama surface wind jet that also produces upwelling, reducing sea surface temperature (SST), increasing sea surface salinity (SSS) and causing an ADT decrease. From June to December, the Choco surface wind jet enhances SST, precipitation and river runoff, which reduces SSS, causing an ADT rise, which in turn forces a weak circulation in the bight, vanishing the PJSC. Interannual variability in the region is strongly affected by El Niño–Southern Oscillation (ENSO); however this climatic variability does not modify the seasonal circulation patterns in the Panama Bight. In contrast, the positive (negative) ENSO phase increases (decreases) SST and ADT in the Panama Bight, with a mean annual difference of 0.9 ∘C and 9.6 cm, respectively, between the two conditions, while its effect on SSS is small. However, as the strong seasonal SST, SSS and ADT ranges are up to 2.2 ∘C, 2.59 g kg−1 and 28.3 cm, the seasonal signal dominates over interannual variations in the bight.

Application of self-organizing maps to AFM-based viscoelastic characterization of breast cancer cell mechanics

Cell mechanical properties have been proposed as label free markers for diagnostic purposes in diseases such as cancer. Cancer cells show altered mechanical phenotypes compared to their healthy counterparts. Atomic Force Microscopy (AFM) is a widely utilized tool to study cell mechanics. These measurements often need skilful users, physical modelling of mechanical properties and expertise in data interpretation. Together with the need to perform many measurements for statistical significance and to probe wide enough areas in tissue structures, the application of machine learning and artificial neural network techniques to automatically classify AFM datasets has received interest recently. We propose the use of self-organizing maps (SOMs) as unsupervised artificial neural network applied to mechanical measurements performed via AFM on epithelial breast cancer cells treated with different substances that affect estrogen receptor signalling. We show changes in mechanical properties due to treatments, as estrogen softened the cells, while resveratrol led to an increase in cell stiffness and viscosity. These data were then used as input for SOMs. Our approach was able to distinguish between estrogen treated, control and resveratrol treated cells in an unsupervised manner. In addition, the maps enabled investigation of the relationship of the input variables.

Visualization-aided multi-criteria decision-making using interpretable self-organizing maps

In multi-criterion optimization, decision-makers (DMs) are not often interested in the complete Pareto-optimal front. Instead, they have preferences favoring specific parts of the front. Multi-criterion decision-making (MCDM) literature provides a plethora of approaches for introducing DM’s preference information in an interactive manner to solve multi-criterion optimization problems. Interactions with DMs can be aided with a user-friendly visualization method or by using special data analysis procedures. An earlier study has indicated the use of self-organizing maps (SOM) as a tool for analyzing Pareto-optimal solutions. In this paper, we demonstrate how a specific MCDM method – NIMBUS – can be executed with the interpretable SOM (iSOM) approach iteratively to arrive at one or more preferred solutions. A visual illustration of the entire high-dimensional search space into multiple reduced two-dimensional spaces allows DMs to have a better understanding of the interactions of the objectives and constraints independently, and execute the NIMBUS decision-making procedure with a more wholistic approach. The paper demonstrates the proposed method on a number of multi- and many-objective numerical and engineering problems. The approach is now ready to be integrated with other popular MCDM methods.

Visualization and Semantic Labeling of Mood States Based on Time-Series Features of Eye Gaze and Facial Expressions by Unsupervised Learning.

This study is intended to develop a stress measurement and visualization system for stress management in terms of simplicity and reliability. We present a classification and visualization method of mood states based on unsupervised machine learning (ML) algorithms. Our proposed method attempts to examine the relation between mood states and extracted categories in human communication from facial expressions, gaze distribution area and density, and rapid eye movements, defined as saccades. Using a psychological check sheet and a communication video with an interlocutor, an original benchmark dataset was obtained from 20 subjects (10 male, 10 female) in their 20s for four or eight weeks at weekly intervals. We used a Profile of Mood States Second edition (POMS2) psychological check sheet to extract total mood disturbance (TMD) and friendliness (F). These two indicators were classified into five categories using self-organizing maps (SOM) and U-Matrix. The relation between gaze and facial expressions was analyzed from the extracted five categories. Data from subjects in the positive categories were found to have a positive correlation with the concentrated distributions of gaze and saccades. Regarding facial expressions, the subjects showed a constant expression time of intentional smiles. By contrast, subjects in negative categories experienced a time difference in intentional smiles. Moreover, three comparative experiment results demonstrated that the feature addition of gaze and facial expressions to TMD and F clarified category boundaries obtained from U-Matrix. We verify that the use of SOM and its two variants is the best combination for the visualization of mood states.

Self-Organizing Maps for Cellular In Silico Staining and Cell Substate Classification.

Cellular composition and structural organization of cells in the tissue determine effective antitumor response and can predict patient outcome and therapy response. Here we present Seg-SOM, a method for dimensionality reduction of cell morphology in H&E-stained tissue images. Seg-SOM resolves cellular tissue heterogeneity and reveals complex tissue architecture. We leverage a self-organizing map (SOM) artificial neural network to group cells based on morphological features like shape and size. Seg-SOM allows for cell segmentation, systematic classification, and in silico cell labeling. We apply the Seg-SOM to a dataset of breast cancer progression images and find that clustering of SOM classes reveals groups of cells corresponding to fibroblasts, epithelial cells, and lymphocytes. We show that labeling the Lymphocyte SOM class on the breast tissue images accurately estimates lymphocytic infiltration. We further demonstrate how to use Seq-SOM in combination with non-negative matrix factorization to statistically describe the interaction of cell subtypes and use the interaction information as highly interpretable features for a histological classifier. Our work provides a framework for use of SOM in human pathology to resolve cellular composition of complex human tissues. We provide a python implementation and an easy-to-use docker deployment, enabling researchers to effortlessly featurize digitalized H&E-stained tissue.

Recognition of Stress Activation by Unobtrusive Multi Sensing Setup

It is recognized that stress conditions play an important role in the definition of individual wellness and represent a major risk factor for most non-communicable diseases. Most studies focus on the evaluation of response to maximal stress conditions while a few of them reports results about the detection/monitoring of response to mild stimulations. In this study, we investigate the capability of some physiological signs and indicators (including Heart Rate, Heart Rate Variability, Respiratory Rate, Galvanic Skin Response) to recognize stress in response to moderate cognitive activation in daily life settings. To achieve this goal, we built up an unobtrusive platform to collect signals from healthy volunteers (10 subjects) undergoing cognitive activation via Stroop Color Word Test. We integrated our dataset with data from the Stress Recognition in the Automobile Drivers dataset. Following data harmonization, signal recordings in both datasets were split into five-minute blocks and a set of 12 features was extracted from each block. A feature selection was implemented by two complementary approaches: Sequential Forward Feature Selection (SFFS) and Auto-Encoder (AE) neural networks. Finally, we explored the use of Self-Organizing Map (SOM) to provide a flexible representation of an individual status. From the initial feature set we have determined, by SFFS analysis, that 2 of them (median Respiratory Rate and number peaks in Galvanic Skin Response signals) can discriminate activation statuses from resting ones. In addition, AE experiments also support that two features can suffice for recognition. Finally, we showed that SOM can provide a comprehensive but compact description of activation statuses allowing a fine prototypical representation of individual status.

A statistical Time-Of-Use tariff based wind resource clustering approach using Self-Organizing Maps

The optimized siting of grid-scale renewable generation is a viable technique to minimize the variable component of the electricity generation portfolio. This process, however, requires simulated meteorological datasets, and consequently, significant computational power to perform detailed studies. This is particularly true for countries with large geographic areas. Clustering is a viable data reduction technique that can be utilized to reduce the computational burden. This work proposes the use of Self-Organizing Maps to partition high-dimensional wind speed data using statistical features derived from Time-Of-Use tariff periods. This approach is undertaken with the view towards the optimization of wind farm siting for grid-support objectives where tariff incentivization is the main driver. The proposed approach is compared with clusters derived using Self-Organizing Maps with the temporal wind speed data for the input feature set. The results show increased cluster granularity, superior validation results and decreased execution time when compared with the temporal clustering approach.

An analytics-based heuristic decomposition of a bilevel multiple-follower cutting stock problem

This paper presents a new class of multiple-follower bilevel problems and a heuristic approach to solving them. In this new class of problems, the followers may be nonlinear, do not share constraints or variables, and are at most weakly constrained. This allows the leader variables to be partitioned among the followers. We show that current approaches for solving multiple-follower problems are unsuitable for our new class of problems and instead we propose a novel analytics-based heuristic decomposition approach. This approach uses Monte Carlo simulation and k-medoids clustering to reduce the bilevel problem to a single level, which can then be solved using integer programming techniques. The examples presented show that our approach produces better solutions and scales up better than the other approaches in the literature. Furthermore, for large problems, we combine our approach with the use of self-organising maps in place of k-medoids clustering, which significantly reduces the clustering times. Finally, we apply our approach to a real-life cutting stock problem. Here a forest harvesting problem is reformulated as a multiple-follower bilevel problem and solved using our approach.

Self-organizing map applied to the choice of internal standards for the determination of Cd, Pb, Sn, and platinum group elements by inductively coupled plasma mass spectrometry

The behaviors of internal standards, according to different flow rates of the cell collision gas (He), were studied for the determination of Cd, Pb, Pd, Pt, Rh, and Sn in samples of fish and mollusks by inductively coupled plasma mass spectrometry (ICP-MS). The elements Bi, Ge, In, Sc, and Y were selected as internal standards, considering their masses and first ionization energies. Addition and recovery experiments were carried out at three concentration levels to evaluate the accuracy of the method applied for the analysis of two samples with different matrices. The results were evaluated using a self-organizing map (SOM). The best analyte/IS pairs were as follows: 114Cd+/74Ge+, 195Pt+/74Ge+, and 208Pb+/74Ge+. For 103Rh+, 106Pd+, and 120Sn+, greater accuracy was achieved without use of an internal standard. Helium gas (2.8 mL min−1) was used in the collision cell for the analytes, except for Sn, and recoveries ranged from 98 to 101% under optimal conditions. The use of SOM as an exploratory analysis tool was an effective approach for selection of the most appropriate internal standards.

Reconstruction of ARNT PAS-B Unfolding Pathways by Steered Molecular Dynamics and Artificial Neural Networks.

Several experimental studies indicated that large conformational changes, including partial domain unfolding, have a role in the functional mechanisms of the basic helix loop helix Per/ARNT/SIM (bHLH-PAS) transcription factors. Recently, single-molecule atomic force microscopy (AFM) revealed two distinct pathways for the mechanical unfolding of the ARNT PAS-B. In this work we used steered molecular dynamics simulations to gain new insights into this process at an atomistic level. To reconstruct and classify pathways sampled in multiple simulations, we designed an original approach based on the use of self-organizing maps (SOMs). This led us to identify two types of unfolding pathways for the ARNT PAS-B, which are in good agreement with the AFM findings. Analysis of average forces mapped on the SOM revealed a stable conformation of the PAS-B along one pathway, which represents a possible structural model for the intermediate state detected by AFM. The approach here proposed will facilitate the study of other signal transmission mechanisms involving the folding/unfolding of PAS domains.

Genetic diversity of weedy rice (Oryza sativa f. spontanea) populations in Sri Lanka: An application of Self Organizing Map (SOM)

Abstract. Weerakoon SR, Somaratne S. 2021. Genetic diversity of weedy rice (Oryza sativa f. spontanea) populations in Sri Lanka: An application of Self Organizing Map (SOM). Asian J Agric 5: 35-43. Weedy rice (WR) (Oryza sativa f. spontanea) has become a major threat in rice cultivation. Discrimination of WR from cultivated rice is difficult since agro-morphology of WR and cultivated rice are overlapping. Molecular markers are useful and can be an informative tool for estimating genetic diversity and relationships in closely related WR eco-types. Self-Organizing Maps (SOM) is an interesting and promising classification tool employing an innovative and data-driven classification method based on unsupervised artificial neural networks. The present study focused on exploring the potential use of SOM to classify WR populations of different eco-climatic zones in Sri Lanka using agro-morphological and molecular data. Separate SOMs for each set of variables, agro-morphological and molecular data were developed. The best SOM was chosen based on the error performance. Findings of SOM analyses showed that certain morphological characters (seedling height, leaf blade width, leaf blade length, culm strength, panicle shattering, seed coat color and leaf angle) and certain molecular characters detected from SSR primers (RM 11, RM 21, RM 14, and RM 280) are important in separation of different WR eco-types satisfactorily. SOM clustering of cultivated, wild, and WR eco-types indicated specific patterns of grouping with respect to climatic conditions of the country. WR eco-types in dry zone and wet zone of the country are closely related to Oryza nivara and O. rufipogon respectively.

Visualizing and Interpreting Unsupervised Solar Wind Classifications

One of the goals of machine learning is to eliminate tedious and arduous repetitive work. The manual and semi-automatic classification of millions of hours of solar wind data from multiple missions can be replaced by automatic algorithms that can discover, in mountains of multi-dimensional data, the real differences in the solar wind properties. In this paper we present how unsupervised clustering techniques can be used to segregate different types of solar wind. We propose the use of advanced data reduction methods to pre-process the data, and we introduce the use of Self-Organizing Maps to visualize and interpret 14 years of ACE data. Finally, we show how these techniques can potentially be used to uncover hidden information, and how they compare with previous manual and automatic categorizations.

Use of data mining in a two‐step process of profiling student preferences in relation to the enhancement of English as a foreign language teaching

AbstractThe paper pursues a twofold goal. The first goal refers to the identification of university students' needs regarding such modifications to English language courses that would improve English as a foreign language (EFL) teaching outcomes. The other goal refers to the methodical issue of achieving the first one. In this aspect, the use of selected data mining techniques in a hierarchical way in real data processing is proposed. These are: (a) Self‐Organizing Map (SOM) dataset segmentation and then (b) market basket analysis applied to the individual SOM segments. The research data were collected from the students' survey concerning their opinion of the EFL teaching process; 347 students of a faculty of a technical university in Poland completed the questionnaire. The use of SOM allowed the identification of homogeneous groups of students, while market basket analysis allowed indicating, within each group, the relationships between student opinions of effective methods of teaching English. In such a way, satisfactory student preference profiles as regards their approach to the improvement of English language competences were developed. On this basis, EFL teaching methods appropriate for the specific profile can be adapted.

Use of self-organizing maps to study sex- and speed-dependent changes in running biomechanics.

Up to 79% of runners get injured every year, with higher rates of injuries occurring in females than males. A self-organizing map (SOM) is a type of artificial neural network that can be used to inspect large datasets and study coordination patterns. The purpose of this study was to use an SOM to study the effects of sex and speed on biomechanical coordination patterns. Thirty-two healthy runners ran on an instrumented treadmill at their long slow distance speed (LSD) and at speed 30% faster (LSD+30%). Vertical ground reaction force (vGRF), vertical tibial acceleration, step parameters, electromyograms (EMG) of six lower limb muscles, and joint angles were collected across speeds. Rate of loading (ROL), tibial impact shock (TIS), coupling angle variability (CAV) and movement pattern proportions for hip/knee sagittal and hip frontal / knee sagittal plane couplings, peak EMG, step length, step rate, and knee and ankle joint angle at initial contact were used as an input for the SOM (37 variables). The analysis identified four clusters (i.e., running patterns). While males and females showed similar distribution across clusters at LSD (p=.36) and at LSD+30% (p=.51), females did exhibit a significant (p=.03) shift between clusters as the speed increased from LSD to LSD+30% whereas males did not (p=.17). The shift was associated with an increase in TIS, ROL, step length, step rate, vastus lateralis EMG, hip flexion/knee extension movement pattern proportion, and a decrease in ST EMG and CAVIC for hip sagittal/knee sagittal coupling. As running speed increased there was a significant change in the coordination pattern in females, which was characterized by increases in several variables that are purported risk factors for running related injuries.

Neural Networks in Accounting: Clustering Firm Performance Using Financial Reporting Data

ABSTRACTThis paper considers the use of neural networks—namely self-organizing maps (SOMs)—to analyze and cluster firms' financial performance. Applying SOMs to financial statement data is a consolidated practice; however, in this paper SOMs are used to overcome several limitations encountered in previous works on financial reporting indicators such as the small number of companies in the sample, the limited number of ratios, the homogeneity of the economic sector, and the lack of explanation and further analysis of the SOM outputs. This study uses a large financial dataset related to more than 3,000 companies belonging to every economic sector; it demonstrates that SOMs can effectively process a large dataset of heterogeneous data. Moreover, the SOM results are supported by detailed explanations of the research methodology applied, and further traditional financial analysis addresses the black box nature of the SOMs and can help professionals in the understanding and use of SOMs.

Assessment and Comparison of Multi-annual Size Profiles of Particulate Matter Monitored at an Urban-industrial Site by an Optical Particle Counter with a Chemometric Approach

Air particulate matter dimensions are a key air quality parameter which can be related to composition, transport properties, exposure and effects on humans and the environment. Optical particle counters are increasingly used for dynamic ambient air particle matter size characterization. Monitoring campaigns lasting several months or years produces millions of single values to be elaborated, requiring effective data treatment procedures for extracting information and knowledge. Data mining algorithms as Self-Organizing Map (SOM) can support exploratory data analysis and pattern recognition in aerosol science. The use of SOM allows to elaborate a high number of data, with powerful visualization features using 2D maps, avoiding to lose information on data variability with data pre-treatments, such as compacting minute data to hourly or daily means. In the present study we describe the optical particle counter data elaboration for particulate matter profile assessment and comparison of a long monitoring time (nearly three years) carried out near residential buildings positioned very close to a steel plant. About twelve millions recorded single values have been handled on the whole. The approach allowed to identify four main particulate matter profiles and follow their variation during the years relating the differences with changes in the plant management and process. The possible applications of the present approach are broad in the field of air quality high frequency long monitoring campaigns with different types of instruments for size and compositional characterization of both particulate matter and gases.

Rummaging through the bin: Modelling marine litter distribution using Artificial Neural Networks.

Marine litter has significant ecological, social and economic impacts, ultimately raising welfare and conservation concerns. Assessing marine litter hotspots or inferring potential areas of accumulation are challenging topics of marine research. Nevertheless, models able to predict the distribution of marine litter on the seabed are still limited. In this work, a set of Artificial Neural Networks were trained to both model the effect of environmental descriptors on litter distribution and estimate the amount of marine litter in the Central Mediterranean Sea. The first goal involved the use of self-organizing maps in order to highlight the importance of environmental descriptors in affecting marine litter density. The second goal was achieved by developing a multilayer perceptron model, which proved to be an efficient method to estimate the regional quantity of seabed marine litter. Results demonstrated that machine learning could be a suitable approach in the assessment of the marine litter issues.