Application Of Self-organizing Maps Research Articles

Self-organizing maps of unbiased ligand-target binding pathways and kinetics.

The interpretation of ligand-target interactions at atomistic resolution is central to most efforts in computational drug discovery and optimization. However, the highly dynamic nature of protein targets, as well as possible induced fit effects, makes difficult to sample many interactions effectively with docking studies or even with large-scale molecular dynamics (MD) simulations. We propose a novel application of Self-Organizing Maps (SOMs) to address the sampling and dynamic mapping tasks, particularly in cases involving ligand flexibility and induced fit. The SOM approach offers a data-driven strategy to create a map of the interaction process and pathways based on unbiased MD. Furthermore, we show how the preliminary SOM mapping is complementary to kinetic analysis, with the employment of both network-based approaches and Markov state models. We demonstrate the method by comprehensively mapping a large dataset of 640 μs of unbiased trajectories sampling the recognition process between the phosphorylated YEEI peptide and its high-specificity target lck-SH2. The integration of SOM into unbiased simulation protocols significantly advances our understanding of the ligand binding mechanism. This approach serves as a potent tool for mapping intricate ligand-target interactions with unprecedented detail, thereby enhancing the characterization of kinetic properties crucial to drug design.

Cognitive profiles across the psychosis continuum

Cognitive impairments are core features in individuals across the psychosis continuum and predict functional outcomes. Nevertheless, substantial variability in cognitive functioning within diagnostic groups, along with considerable overlap with healthy controls, hampers the translation of research findings into personalized treatment planning.Aligned with precision medicine, we employed a data driven machine learning method, self-organizing maps, to conduct transdiagnostic clustering based on cognitive functions in a sample comprising 228 healthy controls, 200 individuals at ultra-high risk for psychosis, and 98 antipsychotic-naïve patients with first-episode psychosis.The self-organizing maps revealed six clinically distinct cognitive profiles that significantly predicted baseline functional level and changes in functional level after one year. Cognitive flexibility in particular, as well as specific executive functions emerged as cardinal in differentiating the profiles. The application of self-organizing maps appears to be a promising approach to inform clinical decision-making based on individualized cognitive profiles, including patient allocation to different interventions. Moreover, this method has the potential to enable cross-diagnostic stratification in research trials, utilizing data-driven subgrouping informed by categories from underlying dimensions of cognition rather than from clinical diagnoses. Finally, the method enables cross-diagnostic profiling across other data modalities, such as brain networks or metabolic subtypes.

Post-pandemic performance of micro, small and medium-sized enterprises: A Self-organizing Maps application

Post-pandemic performance of micro, small and medium-sized enterprises: A Self-organizing Maps application

Application of self-organizing maps to coal elemental data

Understanding the modes of occurrence of elements in coal is important, not only to help properly evaluate the impacts of potentially toxic elements on the environment and human health but also to provide technical guidance for recovering critical elements from coal ash. Statistical and multivariate data analysis methods have widely been used, together with physical and chemical methods, to determine the modes of occurrence of elements in coal. However, some of the statistical methods, e.g., average linkage hierarchical clustering algorithm have some disadvantages (e.g., statistical errors and poor visualization). A self-organizing map is an unsupervised artificial neural network, and it is known for its high data mining capability and excellent data visualization. In contrast to the average linkage hierarchical clustering algorithm that is commonly used for analyzing the modes of occurrence of elements in coal, the self-organizing map algorithm can provide a topological relationship among elements instead of merely providing the groups to which the elements belong. This paper focuses on the application of self-organizing map to coal elemental data for analyzing the modes of occurrence of elements in coal. Samples used in this study are from the Adaohai, Hailiushu, and Datanhao mines, all located in the Daqingshan Coalfield, Inner Mongolia, China. The results obtained from the self-organizing map algorithm are compared with those produced by average linkage hierarchical clustering algorithm. Based on the previous investigations (mainly direct methods) and further analysis, it can be concluded that the results from the self-organizing map algorithm in this investigation are more consistent with the geochemical nature and previous investigations by direct methods than those from average linkage hierarchical clustering algorithm. Consequently, the self-organizing map algorithm is a new reliable and intuitive method for analyzing the modes of occurrence of elements in coal.

On the fast track: Rapid construction of stellar stream paths

ABSTRACT Stellar streams are sensitive probes of the Galactic potential. The likelihood of a stream model given stream data is often assessed using simulations. However, comparing to simulations is challenging when even the stream paths can be hard to quantify. Here we present a novel application of self-organizing maps and first-order Kalman filters to reconstruct a stream’s path, propagating measurement errors and data sparsity into the stream path uncertainty. The technique is Galactic-model independent, non-parametric, and works on phase-wrapped streams. With this technique, we can uniformly analyse and compare data with simulations, enabling both comparison of simulation techniques and ensemble analysis with stream tracks of many stellar streams. Our method is implemented in the public Python package TrackStream, available at https://github.com/nstarman/trackstream.

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.

Self-Organizing Maps in High Energy Physics

The Self-Organizing-Map (SOM) is a widely used neural network for dimensional reduction and clustering. It has yet to find its use in high energy physics. This paper discusses two applications of SOM: first, we find map regions with a high relative content of a rare process (H → WW*). Second we obtain Monte Carlo normalization factors for different physics processes by fitting the dimensionally reduced representation. Analysis and training are performed on ATLAS open data.

ELECTROFACIES DEFINITION AND ZONATION OF THE LOWER CRETACEOUS BARRA VELHA FORMATION CARBONATE RESERVOIR IN THE PRE‐SALT SEQUENCE OF THE SANTOS BASIN, SE BRAZIL

Lower Cretaceous carbonates in the pre‐salt succession in the Santos Basin, eastern Brazil, are highly heterogeneous in terms of their reservoir characteristics as a result of depositional and diagenetic factors. Electrofacies have widely been used for reservoir zonation and, when allied with computer‐based methods such as neural networks, may help with the study of such complex reservoir rocks and with the identification of high‐quality reservoir zones. In this work, an unsupervised artificial neural network known as a self‐organizing map (SOM) was used to carry out a zonation of the pre‐salt carbonates in the Aptian Barra Velha Formation, the main reservoir unit in the Santos Basin. Available data included gramma‐ray, neutron porosity, resistivity deep, sonic, density, photoelectric factor, total porosity and effective porosity profiles from 21 wells together with mineralogical models. Core descriptions and thin section images were used as additional data for the lithological characterization of the electrofacies and consequently for reservoir zonation. A total of four electrofacies were defined from the SOM application, and five reservoir zones were identified.The characterization of the reservoir zones also considered the structural locations of the wells based on the relative depth to top‐ Barra Velha Formation; well locations were classified as structurally high, intermediate or low. Based on the reservoir zone characteristics, the results could be correlated with zonations in previous studies. A general tendency was noted for there to be an increase of finer‐grained sediments in the formation in wells located in structural lows; packstone and mudstone facies were prevalent in these wells and were in general characterized as poor‐quality reservoir rocks. By contrast, the shrubstones and grainstones which were more frequent in structurally high wells comprised higher quality reservoir rocks.The basal reservoir zone showed wide lithological variation compared to the overlying reservoir zones. Grainstone‐dominated facies were identified in the middle of the formation, and the uppermost reservoir zones were characterized by an upward increase in shrubstones and reworked grainstones which in general pointed to better quality reservoirs.

Exploring the Applicability of Self-Organizing Maps for Ecosystem Service Zoning of the Guangdong-Hong Kong-Macao Greater Bay Area

Sustainability is one of the major challenges in the 21st century for humanity. Spatial zoning of ecosystem services is proposed in this study as a solution to meet the demands for the sustainable use of ecosystem services. This study presented a workflow and performed an exploratory analysis using self-organizing maps (SOM) for visualizing the spatial patterns of the ecosystem service value (ESV) of the Guangdong-Hong Kong-Macao Greater Bay Area (GBA). The zoning was performed based on 11 types of ecosystem services, resulting in 11 ecosystem service zones. Each of the zones derived has its unique characteristics in terms of the dominating ecosystem service types, ESV, land use/land cover patterns, and associated human activity levels. It is recommended that reasonable and effective utilization of the ecosystem services in the GBA should be based on its zonal characteristics rather than haphazard exploitations, which can contribute to the sustainable economy and environment of the region. The applicability of SOM for the GBA ecosystem service zoning has been demonstrated in this study. However, it should be stressed that the method and workflow presented in this study should mainly be used for supporting decision-making rather than used for deriving gold-standard zoning maps.

Application of self-organizing maps and K–Means methods to classify summer heat wave weather patterns in Viet Nam

Application of self-organizing maps and K–Means methods to classify summer heat wave weather patterns in Viet Nam

Application of self-organizing map (SOM)-based approach to explore the relationship between land use and water quality in Deqing County, Taihu Lake Basin

Characterizing the spatial patterns of surface water quality is of great significance for water resources management and protection. A great number of studies have addressed the fact that water quality is often subjected to surrounding land use patterns, but the relationship between them remains obscure. In this study, we develop a self-organizing map (SOM)-based approach to explore the relationship between land use and water quality in Deqing County, Zhejiang Province, which locates in the hinterland of the Yangtze River Delta in the Taihu Lake Basin. Water samples were collected from 163 streams, ponds or lakes and then clustered into six groups based on five water quality variables: pH, ammonium-N, total phosphate, dissolved oxygen and permanganate index using the SOM approach. The proportion of eight land use categories, which are forest, cropland, shrubland, grassland, water, wetland, impervious land and bare land were associated with water quality to explain its spatial variation. The altitude decreases from west to east and the eastern part of Deqing County is characterized by floodplain and streams network, which makes watershed delineating real hard work. Therefore polder and multiple buffer scales were considered to carry out the spatial discretization process. For land use parameters, sub-watershed scale was surpassing buffer scale in the elevated area, but only buffer scale can be applied in the whole area. Polders totally failed to reveal the relationship between land use and water quality. Our results demonstrate whether point source (PS) or non-point source (NPS) pollution exerts greater influence on water quality in different clusters. The methods and the results demonstrated in this research can help local government improve water resources and land use management strategies practically.

Target-scale prospectivity modeling for gold mineralization within the Rajapalot Au-Co project area in northern Fennoscandian Shield, Finland. Part 2: Application of self-organizing maps and artificial neural networks for exploration targeting

This paper is part of a two-publication series reporting target-scale prospectivity modeling results for gold enrichment in the Rajapalot project area. The study area is located in the Northern Fennoscandian Shield in Finland and consists of high-grade epigenetic-hydrothermal Au-Co prospects. The first publication of this series described the implementation of knowledge-driven- and hybrid- expert systems. In this second paper we demonstrate the application of machine learning methods such as unsupervised self-organizing maps (SOM) and K-means clustering and supervised artificial neural networks (ANNs). The results from SOM allowed the mapping of deposit-related geological patterns in the evidential layers. K-means clustering of the SOM results identified data clusters favorable for gold enrichment. Quantitative prospectivity values were computed using an ANN model trained on the SOM codebook vectors. This improved the resolution of the prospective-exploration areas mapped within the geospatial domains of the clusters, and reduced the exploration search areas. The ANN result has a significant area under curve value of 0.869 in the receiver operating characteristic plots, confirming the predictive ability of the model. Results identify new areas for exploration, as well as help prioritise drilling targets around the currently less explored prospects. In this second paper we therefore conclude that (1) SOM is an effective method for mapping and visualization of patterns (in the evidential layers) related to mineral deposits at the target scale, and (2) supervised classification of the SOM results, such as using an ANN, can help define new target areas for drilling around the existing prospects. Finally, this two-publication series collectively demonstrates that advanced machine-learning based pattern recognition and prospectivity modeling routines can be used for exploration targeting at camp- and mine- scales. The knowledge-driven methods are efficient at modeling mineralization processes and the data-driven methods are effective at identifying geological attributes related to mineral deposits and occurrences.

Galaxy and mass assembly (GAMA): Self-Organizing Map application on nearby galaxies

ABSTRACTGalaxy populations show bimodality in a variety of properties: stellar mass, colour, specific star-formation rate, size, and Sérsic index. These parameters are our feature space. We use an existing sample of 7556 galaxies from the Galaxy and Mass Assembly (GAMA) survey, represented using five features and the K-means clustering technique, showed that the bimodalities are the manifestation of a more complex population structure, represented by between two and six clusters. Here we use Self-Organizing Maps (SOM), an unsupervised learning technique that can be used to visualize similarity in a higher dimensional space using a 2D representation, to map these 5D clusters in the feature space on to 2D projections. To further analyse these clusters, using the SOM information, we agree with previous results that the sub-populations found in the feature space can be reasonably mapped on to three or five clusters. We explore where the ‘green valley’ galaxies are mapped on to the SOM, indicating multiple interstitial populations within the green valley population. Finally, we use the projection of the SOM to verify whether morphological information provided by GalaxyZoo users, for example, if features are visible, can be mapped on to the SOM-generated map. Voting on whether galaxies are smooth, likely ellipticals, or ‘featured’ can reasonably be separated but smaller morphological features (bar, spiral arms) can not. SOMs promise to be a useful tool to map and identify instructive sub-populations in multidimensional galaxy survey feature space, provided they are large enough.

3D Simulation and Artificial Neural Networks Application in the Diffusion Study of Champignon Mushroom/Solution Interface During the Salting

The influence of the film formed during the salting of champignon mushrooms with brine containing NaCl and KCl was modeled using the finite elements method (FEM). It was verified that the film formed on the mushroom surface had a greater influence in the static salting since the diffusion of the ions was 7.5-fold smaller in this system than in the stirred salting. The application of self-organizing maps showed that the ions diffusion along the surface of the solid presented a heterogeneous occurrence and depended on the region for both static and stirred salting. A direct relation was observed among the mushroom surface morphology, the salts diffusion behavior, and the film formation. In addition, the film was not completely extinguished in the stirred system, although it has a minimal influence as the film formation is also dependent on the biosolid surface.

ERRATUM: Application of self-organizing maps to evaluate the influence and behavior of the film formed during salting of Prato cheese

ERRATUM: Application of self-organizing maps to evaluate the influence and behavior of the film formed during salting of Prato cheese

Application of self-organizing map (SOM) and K-means clustering algorithms for portraying geochemical anomaly patterns in Moalleman district, NE Iran

In this paper, in order to reveal the regional geochemical patterns of regularly sampled stream sediment data, we have employed the K-means and self-organizing map (SOM) as clustering methods in the Moalleman district, northeast Iran. Initially, a set of analyzed elements of geochemical data was subjected to isometric log-ratio (ilr) transformation to address the closure problem related to geochemical data, then, ordinary principal component analysis (PCA) was utilized for recognizing the internal relations between selected elements (As, Au, Cu, Pb, Sb and Zn). Subsequently, the K-means and SOM as unsupervised clustering methods were applied based on PC1 (Cu-Pb-Zn aggregation) and PC2 (Au-As-Sb aggregation) to distinguish different populations of multi-element geochemical indicators. In this regard, Silhouette Width (SW) was implemented for computing the optimal cluster number in K-means clustering method. In the next step, due to the presence of numerous copper mineral deposits/occurrences in the study area, we opted to implement the supervised SOM on ilr-transformed values of Cu-Pb-Zn elements for delineating high anomalous zones. For this purpose, a confusion matrix based on training and out-of-bag (OOB) data was developed for the supervised SOM model and the results indicated the accuracy of 96.27% and 94.26%, respectively. Moreover, success-rate curves were used for assessing the overall performance of K-means and SOM (unsupervised and supervised) models. Experimental outcomes represented the superiority of SOM models (especially the supervised SOM) over K-means in delineating the geochemical anomaly targets which can be used as an effective and powerful tool for discovering the complex patterns among variables in exploratory geochemical data.

Trends and projection of heavy snowfall in Hokkaido, Japan as an application of self-organizing map

AbstractThis paper showed the frequency of local-scale heavy winter snowfall in Hokkaido, Japan, its historical change, and its response to global warming using self-organizing map (SOM) of synoptic-scale sea-level pressure anomaly. Heavy snowfall days were here defined as days when the snowfall exceeded 10 mm in water equivalent. It was shown that the SOMs can be grouped into three categories for heavy snowfall days: 1) a passage of extratropical cyclones to the south of Hokkaido, 2) a pressure pattern between the Siberian high and the Aleutian low, and 3) a low-pressure anomaly just to the east of Hokkaido. Groups 1 and 2 were associated with heavy snowfall in Hiroo (located in southeastern Hokkaido) and in Iwamizawa (western Hokkaido), respectively, and heavy snowfall in Sapporo (western Hokkaido) was related to Group 3. The large-ensemble historical simulation reproduced the observed increasing trend in Group 2 and future projection revealed that Group 2 was related to a negative phase of the Western Pacific pattern and the frequency of this group would increase in the future. Heavy snowfall days associated with SOM Group 2 would also increase due to the increase in water vapor and preferable weather patterns in global warming climate, in contrast to the decrease of heavy snowfall days in other sites associated with SOM Group 1.

Study on Land Consolidation Zoning in Hubei Province Based on the Coupling of Neural Network and Cluster Analysis

Land consolidation zoning is the basis of land reclamation planning and provides a precondition for land management work. Zoning is beneficial to scientific decision-making regarding the use of cultivated land resources, and helps to ensure the quality and production level of cultivated land. Hence, land consolidation zoning is of national significance in terms of safeguarding food security. In this paper, an indicator system for land evaluation and zoning is described based on indicators of natural ecological conditions, socioeconomic conditions, land-use conditions, and the land consolidation potential of Hubei province. The application of self-organizing maps and Hierarchical clustering results in 11 and 12 categories, respectively. After comparing and adjusting these results according to the real situation and key points in the land consolidation of each zone, seven integrated categories are determined. Finally, the land consolidation zoning scheme for these seven integrated partitions is shown to be suitable for Hubei province and in line with the real situation. The proposed scheme offers support for land remediation in a coordinated and orderly way, and provides a scientific basis for improving land utilization efficiency and grain production capacity in Hubei province.

A Bernoulli-Gamma hierarchical Bayesian model for daily rainfall forecasts

We consider stochastic weather models originally developed for rainfall simulations to build a hierarchical Bayesian mixture model for daily rainfall forecasts using endogenous and external information. We model daily rainfall as a seasonal-varying mixture of a Bernoulli distribution for rainfall occurrence and a gamma distribution for the rainfall amount. The model scheme allows the inclusion of predictors to reduce the bias and variance of the forecasts, while the hierarchical Bayesian framework promotes a better understanding and reduction in parameter uncertainties, especially for gauges with short records, as well as supports the estimation of regional parameters that could be employed for forecasts at ungauged sites. The model was tested using 47 years (1973–2019) of daily rainfall data from 60 gauges in South Korea. Climate indices derived from the low-level wind over the region were analyzed using Principal Component Analysis (PCA) and embodied into the model to enhance its forecast skills. The model structure was based on a detailed exploratory data analysis, which included the application of Self-Organizing Maps (SOM) to examine the spatio-temporal patterns of rainfall. Cross-validated results reveal improved skills over reference models based on climatology and persistence up to a three days lead time. The average gains in metrics such as the Brier and Winkler skill scores vary from 5% to 50%, while the average correlation skill between predictions and observations reach values up to 0.55. The gains beyond a three days lead time are marginal, but the underlying structure of the proposed model still encourages its use over the reference models, being a step forward in improving real-time daily rainfall forecasts for the region. It has also a great potential to be combined with weather model forecasts and applied in other places across the world.

Supervised mineral exploration targeting and the challenges with the selection of deposit and non-deposit sites thereof

Selection of non-deposit sites is a challenging issue affecting the application of supervised algorithms for modeling mineral exploration targets. For this, equal number of deposit and non-deposit sites has been widely applied for training purposes. In this paper, we investigated the effect of changes in the number of non-deposit sites on the effectiveness of exploration targeting models while the number of deposit sites is constant. The results obtained demonstrated that exploration targeting models are affected by the ratio of non-deposit and deposit sites. Thus, balancing between the number of deposit and non-deposit sites is an efficient way to produce more-reliable exploration targets when supervised algorithms are applied for modeling. The idea of this research came from the fact that mineralization is a rare event, and therefore, in a region of interest number of non-deposit sites is much more than that of deposit events. To illustrate the procedure proposed, we used an exploration dataset of porphyry Cu mineralization in Chahargonbad area, SE Iran. A sequence application of self-organizing map and multilayer perceptron neural network algorithm was applied to better illustration of the changing effects of the number of non-deposit sites on the ensuing exploration targeting models.