SOM Clustering Research Articles

Intelligent treatment system based on bioinformatics and neuro-immune-digestive tract diseases

There are currently many clinical issues plaguing digestive oncologists, including the impact of dietary habits on the occurrence of early gastrointestinal cancer, the association between irritable bowel syndrome, ulcerative colitis, and early gastrointestinal cancer, the expected changes in gastrointestinal cancer after IBS, IBD, and susceptibility gene mutations, and the mechanisms by which susceptibility gene mutations in gastrointestinal cancer lead to different diseases. Meanwhile, as a carcinogen, betel nut is also prone to cancer and emotional disorders, and there are also many doctor-patient conflicts in the current clinical treatment of digestive tract tumors. To address the aforementioned issues, the research team analyzed dietary habits data using SOM clustering algorithm and isolated forest algorithm. Based on bioinformatics, new therapies have been developed for the pathogenesis and immune pathways of various cancers in the digestive system. Using network pharmacology methods, explore the toxicological mechanism of betel nut and its mechanism in mediating emotional disorders. Molecular docking is used to elucidate drug prediction. Based on four AI algorithms, the occurrence of gastrointestinal cancer was predicted through dietary habits. Using computer vision technology to address the deficiencies in early clinical reporting of cancer. This article provides clinical doctors with new treatment tools.

Evaluation and application of local fiscal science and technology expenditure efficiency in the Yangtze River Delta urban agglomeration based on AHP-TOPSIS and SOM clustering

This study selected 18 indicators from four dimensions: the scale of regional fiscal science and technology expenditure, the social impact of science and technology, the efficiency of science and technology expenditure, and the environment for scientific and technological innovation, and constructed an evaluation index system for the efficiency of local fiscal science and technology expenditure in the Yangtze River Delta urban agglomeration. The AHP-TOPSIS model was used to determine the weights of evaluation indicators, and the efficiency ranking of fiscal science and technology expenditures in various cities of the Yangtze River Delta urban agglomeration was quantified. Furthermore, using the SOM algorithm for clustering analysis, the level division of innovation capability in the Yangtze River Delta urban agglomeration was clarified. This study provides a scientific tool and method for evaluating the efficiency of local financial science and technology expenditure in the Yangtze River Delta urban agglomeration, and provides solid theoretical support and practical guidance for promoting the coordinated development of regional science and technology innovation.

Analysis of power plant outage event results based on SOM clustering

Through the file mining analysis of fault event analysis reports, it is possible to discover the fault mode characteristics that exist between different fault events, thereby providing empirical guidance for safety management. The self-organizing mapping SOM clustering method was used to analyze the fault cause characteristics of 520 unplanned shutdown events in power plants. The main influencing factors are divided into six primary indicators: external factors, technical factors, equipment failures, pipeline failures, logical factors, and human factors, and 36 secondary indicators of influencing factors. Based on different indicator characteristics, the power plant shutdown event can be mapped into a two-dimensional fault feature event map, which can intuitively obtain the visualization results of shutdown events with different influencing factor dimensions. Through SOM clustering analysis, the shutdown fault events were ultimately divided into three different clusters. In cluster 1, the highest proportion is ultimately caused by external influencing factors and equipment failures. In cluster 2, it is mainly affected by human factors and control logic factors. In cluster 3, it is mainly the pipeline aging failure caused by long-term running time factors. By visualizing the results of intuitive fault characteristics, it can provide diversified knowledge information for risk management and technical personnel to take corresponding measures to improve the safe and stable operation of the unit.

Research on Regional Green Manufacturing Development Level Measurement Based on Temporal and Spatial Entropy Weighted TOPSIS and SOM Clustering in Cloud Manufacturing

Research on Regional Green Manufacturing Development Level Measurement Based on Temporal and Spatial Entropy Weighted TOPSIS and SOM Clustering in Cloud Manufacturing

No Reference Image Quality Assessment Based on DCT and SOM Clustering

No Reference Image Quality Assessment Based on DCT and SOM Clustering

A model for identifying glass artifacts based on SOM clustering analysis and random forest algorithm

With the deepening of human awareness of heritage conservation today, the category identification of excavated glass artifacts is particularly important. This paper draws on data related to the chemical composition of some glass artifacts. The data were analyzed by Mann-Whitney U analysis, SOM cluster analysis, and random forest algorithm, and a model was developed to accurately identify the categories of glass artifacts based on their chemical composition. The results show that the three chemical components of potassium oxide, barium oxide, and lead oxide have the greatest influence on the weathering and corrosion of glass artifacts through Mann-Whitney U analysis; the SOM cluster analysis model shows that the glass artifacts with high potassium can be divided into two subclasses, and the glass artifacts with lead and barium can be divided into three subclasses. Finally, we combined the results of the existing analysis and used the random forest algorithm to establish a model for accurate identification of glass artifacts based on their chemical composition. The sensitivity test shows that the model has high robustness and accuracy. This method will play an important role in the accurate identification of glass artifacts of unknown categories in the future.

Feature selection for binary classification based on class labeling, SOM, and hierarchical clustering

Feature selection plays an important role in algorithms for processing high-dimensional data. Traditional pattern classification and information theory methods are widely applied to feature selection methods. However, traditional pattern classification methods such as Fisher Score, Laplacian Score, and relief use class labels inadequately. Previous information theory based feature selection methods such as MIFS ignore the intra-class to tight inter-class to sparse property of the samples. To address these problems, a feature selection algorithm for the binary classification problem is proposed, which is based on class label transformation using self-organizing mapping neural network (SOM) and cohesive hierarchical clustering. The algorithm first converts class labels without numerical meaning into numerical values that can participate in operations and retain classification information through class label mapping, and constitutes a two-dimensional vector from it and the attribute values to be judged. Then, these two-dimensional vectors are clustered by using SOM neural network and hierarchical clustering. Finally, evaluation function value is calculated, that is closely related to intra-cluster to tightness, inter-cluster separation, and division accuracy after clustering, and is used to evaluate the ability of alternative attributes to distinguish between classes. It is experimentally verified that the algorithm is robust and can effectively screen attributes with strong classification ability and improve the prediction performance of the classifier.

Diversity of Catechin Content in the Leaves of Persimmon Germplasms

Catechins are essential phenolic compounds in persimmon. However, the catechin content in the leaves of persimmon germplasms has not been systematically and comprehensively evaluated. To systematically assess catechin variation in different growth stages and astringency types of persimmon leaves, the young and adult leaves catechin content of 249 persimmon germplasms from Korea, Japan, and 16 provinces in China was quantified using an HPLC method. The results showed that the content of (−)-epicatechin (EC) and (−)-gallocatechin gallate (GCG) had higher coefficient of variation (CV) values in persimmon young leaves (106.74%) and adult leaves (91.49%), respectively. The genetic diversity H’ of (+)-catechin (CA) and (+)-gallocatechin (GC) content was higher in young leaves (2.00 and 2.03), and the H’ value of (−)-epigallocatechin (EGC) and CA content was higher in adult leaves (1.98 and 1.92). The non-galloylated catechins, EGC, EC, and CA, were significantly positively correlated with each other in both young and adult leaves. Catechins of persimmon leaves showed different accumulation patterns in different growth stages and astringency types; (−)-epigallocatechin gallate (EGCG) and CA had the highest average content in the persimmon germplasms of Guangxi and Shanxi provinces of China, respectively. The content of (+)-catechin gallate (CG) was found to be significantly higher in Japanese pollination-constant non-astringent (J-PCNA) persimmon leaves, and EC, GCG had the highest levels in Chinese PCNA (C-PCNA) type. The 249 germplasms were classified into five clusters (Cluster I-V) by SOM clustering based on the content of nine catechins, with EC, CA, and GC having the highest content in Cluster V of the adult leaves. In addition, four excellent germplasms of Cluster III had the highest EGC and GC content in young leaves. Thus, the critical period for catechin utilization of persimmon leaves can be further determined, and provide theoretical references for excellent germplasm screening.

Measuring Urban Poverty Spatial by Remote Sensing and Social Sensing Data: A Fine-Scale Empirical Study from Zhengzhou

Urban poverty is a major obstacle to the healthy development of urbanization. Identifying and mapping urban poverty is of great significance to sustainable urban development. Traditional data and methods cannot measure urban poverty at a fine scale. Besides, existing studies often ignore the impact of the built environment and fail to consider the equal importance of poverty indicators. The emerging multi-source big data provide new opportunities for accurately measuring and monitoring urban poverty. This study aims to map urban poverty spatial at a fine scale by using multi-source big data, including social sensing and remote sensing data. The urban core of Zhengzhou is selected as the study area. The characteristics of the community’s living environment are quantified by accessibility, block vitality, per unit rent, public service infrastructure, and socio-economic factors. The urban poverty spatial index (SI) model is constructed by using the multiplier index of the factors. The SOM clustering method is employed to identify urban poverty space based on the developed SI. The performance of the proposed SI model is evaluated at the neighborhood scale. The results show that the urban poverty spatial measurement method based on multi-source big data can capture spatial patterns of typical urban poverty with relatively high accuracy. Compared with the urban poverty space measured based on remote sensing data, it considers the built environment and socio-economic factors in the identification of the inner city poverty space, and avoids being affected by the texture information of the physical surface of the residential area and the external structure of the buildings. Overall, this study can provide a comprehensive, cost-effective, and efficient method for the refined management of urban poverty space and the improvement of built environment quality.

Color segmentation of multicolor porous printed fabrics by conjugating SOM and EDSC clustering algorithms

In order to achieve accurate color segmentation of printed fabrics, the color segmentation algorithm combining the self-organizing maps neural network and the efficient dense subspace clustering was proposed in this paper. After pre-processing of the fabric image, the primary clustering was implemented by the self-organizing maps algorithm, then the secondary clustering was done by the efficient dense subspace clustering algorithm. The optimal silhouette coefficient is introduced into the clustering process of the efficient dense subspace clustering algorithm to determine the number of clustering centers automatically. Finally, by the post-processing including gray-scale transformation, binarization and open operation, the mis-segmentation of edge color was eliminated, making the algorithm more suitable for industrial application. Experiments were carried out and results show that the algorithm proposed in this paper can recognize the color of small areas accurately and segment the color of complex printed fabric images. The color segmented results of 20 printed fabrics show that the accuracy of the algorithm proposed in this paper reaches 88.3%.

Modeling customer satisfaction through online reviews: A FlowSort group decision model under probabilistic linguistic settings

Modeling customer satisfaction from online customer reviews (OCRs) has become a practical issue. OCRs provide enough information to aid service providers in measuring customer satisfaction. However, the voluminous nature and the ambiguity surrounding OCRs increase the difficulty for service providers to measure customer satisfaction. Since different customers have diverse perceptions, priorities, and preferences, it is appropriate to measure customer satisfaction from customer segmentation. Thus, we establish a probabilistic linguistic group decision (PLGD)-FlowSort methodology to model customer satisfaction. This methodology employs the Latent Dirichlet Allocation (LDA) topic model to extract the relevant customer satisfaction dimensions (CSDs) from OCRs. Then, based on the SOM clustering technique, we segment customers based on their satisfaction degree. To analyze OCRs, we compute the sentiment scores of the reviews using an unsupervised machine learning algorithm and convert the sentiment scores into probabilistic linguistic term sets (PLTSs). To objectively determine the weight information of the segments and the CSDs, we design the probabilistic linguistic (PL)-projection method and the probabilistic linguistic-correlation coefficient standard deviation (PL-CCSD) method, respectively. The PLGD-FlowSort is developed to measure customer satisfaction towards several service providers and categorize them into different satisfaction levels. Finally, we apply the proposed methodology to a case of measuring customer satisfaction towards mobile payments in Ghana and further perform a comparative analysis to test the robustness of our work.

Genetic Landscape of Relapsed and Refractory Diffuse Large B-Cell Lymphoma: A Systemic Review and Association Analysis With Next-Generation Sequencing.

In our research, we screened 1,495 documents, compiled the whole-exome sequencing data of several studies, formed a data set including 92 observations of RRDLBCL (Relapsed and refractory diffuse large B-cell lymphoma), and performed association analysis on the high-frequency mutations among them. The most common mutations in the data set include TTN, KMT2D, TP53, IGLL5, CREBBP, BCL2, MYD88, and SOCS1 etc. Among these, CREBBP, KMT2D, and BCL2 have a strong association with each other, and SOCS1 has a strong association with genes such as STAT6, ACTB, CIITA, ITPKB, and GNA13. TP53 lacks significant associations with most genes. Through SOM clustering, expression-level analysis and protein interaction analysis of common gene mutations, we believe that RRDLBCL can be divided into five main types. We tested the function of the model and described the clinical characteristics of each subtype through a targeted sequencing RRDLBCL cohort of 96 patients. The classification is stated as follows: 1) JAK-STAT-related type: including STAT6, SOCS1, CIITA, etc. The genetic lineage is similar to PMBL and cHL. Retrospective analysis suggests that this subtype responds poorly to induction therapy (R-CHOP, p < 0.05). 2) BCL-CREBBP type: Epigenetic mutations such as KMT2D and CREBBP are more common in this type, and are often accompanied by BCL2 and EZH2 mutations. 3) MCD type: including MYD88 and CD79B, PIM1 is more common in this subtype. 4) TP53 mutation: TP53 mutant patients, which suggests the worst prognosis (p < 0.05) and worst response to CART treatment. 5) Undefined type (Sparse item type): Major Genetic Change Lacking Type, which has a better prognosis and better response to CART treatment. We also reviewed the literature from recent years concerning the previously mentioned common gene mutations.

Computational Design of Indoor Arenas (CDIA): Integrating multi-functional spaces and long-span roof structures

Indoor arenas are important public buildings catering for various activities (e.g., sports events, stage performances, assemblies, exhibitions, and daily sports for the public) and serving as landmarks in urban contexts. The multi-functional space and long-span roof structure of an indoor arena are highly interrelated, which impact the multi-functionality and structural performance and mainly define the overall form of the building. Therefore, it is crucial to integrate the multi-functional space and long-span roof structure to formulate proper forms for indoor arenas, in order to satisfy various design requirements during the conceptual design. This thesis aims at formulating a computational design method, ‘Computational Design of Indoor Arena (CDIA)’, to support the conceptual design of indoor arenas by using the computational techniques of parametric modelling, Building Performance Simulations (BPSs), Multi-Objective Optimizations (MOOs), surrogate models based on Multi-Layer Perceptron Neural Network (MLPNN), and clustering based on Self-Organizing Map (SOM clustering). In the formulation of CDIA, these techniques are modified, improved and organized into five components and three workflows, to satisfy the demands of the conceptual design of indoor arenas.

DISCOVERING HIDDEN CLUSTER STRUCTURES IN CITIZEN COMPLAINT CALL VIA SOM AND ASSOCIATION RULE TECHNIQUE

Significant revolution in different organizations chief’s point of view toward customer treating and the level of product presentation or services resulted in redefining the structure of these organizations based on this point of view. The municipal services are very important as well. The strategy of “CRM” which was so successful in the private sector and has been applying as “CiRM” in the public sector of developed countries could be very useful for this achievement. The main goal of citizen management is realizing the citizen’s needs and demands, improving communication through connection with citizens and optimizing it to increase the level of their satisfaction. The government agencies do it based on their idea and point of view cause the citizen are valuable assets in the planning of services and reduction of costs. This study proposes a combined data mining method to discover hidden knowledge in call citizen compliant of the municipality of Tehran. A Self-organizing map neural network was used to identifying and classifying citizen needs based on RFM analysis. It also classified citizen needs into three majors. the result of classification and clustering of SOM has created a new feature to profiled call’s customer to identify temporal-spatial patterns of problems by using an association rule with the Apriori algorithm. The results of this idea demonstrate that accordance of citizens call compliant in a different area and discovering hidden knowledge can facilitate the performance of human recourse in improving services to citizens.

DM-pruning CADJ graphs for SOM clustering

As topology representing networks, the Cumulative ADJacency graph CADJ and its symmetric version $${{\text{CONN}}} = {\text{CADJ}} + {\text{CADJ}}^T$$ Tasdemir and Merenyi (IEEE Trans Neural Netw 20(4): 549–562, 2009), can be utilized as inputs to graph-based clustering (GBC) paradigms for partitioning the learned prototypes of a vector quantizer. To express complex data faithfully, CADJ must typically be pruned (thresholded, or made sparse) to be most effective as an input to GBC routines, whether they be algorithmic or driven by human assessment. This work, given in two parts, develops a formal framework for CADJ pruning as a preprocessing (sparsifying) step to improve CADJ’s use in any GBC routine. That is, rather than advocating a particular GBC method, our goal is development of sensible logic for creating sparse CADJ inputs to the entire family of GBC methods. Part 1 defines an overall quality measure for each CADJ edge by extending lines of reasoning used successfully in the past to prune CONN graphs. Part 2 introduces a Bayesian Dirichlet-multinomial (DM) model of CADJ edge weights with an intelligent prior constructed through analysis of the Voronoi tessellation generated by the vector quantization. The DM likelihood offers an internal assessment of information loss resulting from iterative CADJ edge removal, which is used to determine an optimal stopping criterion for the pruning process. We show that DM-Pruned CADJ graphs lead to GBCs comparable to the best previously achieved on highly structured real data.

Unsupervised machine learning using 3D seismic data applied to reservoir evaluation and rock type identification

Net reservoir discrimination and rock type identification play vital roles in determining reservoir quality, distribution, and identification of stratigraphic baffles for optimizing drilling plans and economic petroleum recovery. Although it is challenging to discriminate small changes in reservoir properties or identify thin stratigraphic barriers below seismic resolution from conventional seismic amplitude data, we have found that seismic attributes aid in defining the reservoir architecture, properties, and stratigraphic baffles. However, analyzing numerous individual attributes is a time-consuming process and may have limitations for revealing small petrophysical changes within a reservoir. Using the Maui 3D seismic data acquired in offshore Taranaki Basin, New Zealand, we generate typical instantaneous and spectral decomposition seismic attributes that are sensitive to lithologic variations and changes in reservoir properties. Using the most common petrophysical and rock typing classification methods, the rock quality and heterogeneity of the C1 Sand reservoir are studied for four wells located within the 3D seismic volume. We find that integrating the geologic content of a combination of eight spectral instantaneous attribute volumes using an unsupervised machine-learning algorithm (self-organizing maps [SOMs]) results in a classification volume that can highlight reservoir distribution and identify stratigraphic baffles by correlating the SOM clusters with discrete net reservoir and flow-unit logs. We find that SOM classification of natural clusters of multiattribute samples in the attribute space is sensitive to subtle changes within the reservoir’s petrophysical properties. We find that SOM clusters appear to be more sensitive to porosity variations compared with lithologic changes within the reservoir. Thus, this method helps us to understand reservoir quality and heterogeneity in addition to illuminating thin reservoirs and stratigraphic baffles.

Regional classification of total suspended matter in coastal areas of South Korea

Total suspended matter (TSM) in coastal areas is affected by various sediment transport mechanisms. Thus, identifying systematically and robustly the regional pattern of TSM data is of great use for the management of coastal areas. The TSM sampling stations have been regionally clustered using an iterative toroidal SOM–k–means approach, which overcomes the cluster number determination and local optima problems. The iterative toroidal SOM–k–means is a two-step clustering approach that finds an optimal clustering result based on DBI by repeating the toroidal SOM and k-means clustering. Six variables (longitude, latitude, surface TSM in February and August, bottom TSM in February and August) were used as clustering variables, and cluster analyses were conducted while varying combinations of variables. All clustering results were analyzed using the sediment distribution, ocean currents, and tidal currents around South Korean coastal areas. Additionally, tidal flats around the Yeomha Channel, Hampyeong Bay, and Doam Bay, which are associated with sediment dispersion of turbid water, bay erosion, and tidal forcing, respectively, were identified during clustering analyses. This study's outcome provides the basis of regional sediment transport analysis in South Korea and an example of physical analysis based on the data-driven approach.

The Use of 3D Convolutional Autoencoder in Fault and Fracture Network Characterization

Conventional pattern recognition methods directly use 1D poststack data or 2D prestack data for the statistical pattern recognition of fault and fracture network, thereby ignoring the spatial structure information in 3D seismic data. As a result, the generated fault and fracture network is not distinguishable and has poor continuity. In this paper, a fault and fracture network characterization method based on 3D convolutional autoencoder is proposed. First, in the autoencoder training frame, 3D prestack data are used as input, and the 3D convolution operation is used to mine the spatial structure information to the maximum and gradually reduce the spatial dimension of the input. Then, the residual network is used to recover the input’s details and the corresponding spatial dimension. Lastly, the hidden features extracted by the encoders are recognized via k -means, SOM, and two-step clustering analysis. The validity of the method is verified by testing the seismic simulation data and applying real seismic data. The 3D convolution can directly process the seismic data and maximize the prestack texture attributes and spatial structure information provided by 3D seismic data without dimensionality reduction and other preprocessing operations. The interleaving convolution layer and residual block overcome low learning and accuracy rates due to the deepening of networks.

Computational Design of Indoor Arenas (CDIA)

Indoor arenas are important public buildings catering for various activities (e.g., sports events, stage performances, assemblies, exhibitions, and daily sports for the public) and serving as landmarks in urban contexts. The multi-functional space and long-span roof structure of an indoor arena are highly interrelated, which impact the multi-functionality and structural performance and mainly define the overall form of the building. Therefore, it is crucial to integrate the multi-functional space and long-span roof structure to formulate proper forms for indoor arenas, in order to satisfy various design requirements during the conceptual design. This thesis aims at formulating a computational design method, ‘Computational Design of Indoor Arena (CDIA)’, to support the conceptual design of indoor arenas by using the computational techniques of parametric modelling, Building Performance Simulations (BPSs), Multi-Objective Optimizations (MOOs), surrogate models based on Multi-Layer Perceptron Neural Network (MLPNN), and clustering based on Self-Organizing Map (SOM clustering). In the formulation of CDIA, these techniques are modified, improved and organized into five components and three workflows, to satisfy the demands of the conceptual design of indoor arenas.

Computational Design of Indoor Arenas (CDIA)

Indoor arenas are important public buildings catering for various activities (e.g., sports events, stage performances, assemblies, exhibitions, and daily sports for the public) and serving as landmarks in urban contexts. The multi-functional space and long-span roof structure of an indoor arena are highly interrelated, which impact the multi-functionality and structural performance and mainly define the overall form of the building. Therefore, it is crucial to integrate the multi-functional space and long-span roof structure to formulate proper forms for indoor arenas, in order to satisfy various design requirements during the conceptual design. This thesis aims at formulating a computational design method, ‘Computational Design of Indoor Arena (CDIA)’, to support the conceptual design of indoor arenas by using the computational techniques of parametric modelling, Building Performance Simulations (BPSs), Multi-Objective Optimizations (MOOs), surrogate models based on Multi-Layer Perceptron Neural Network (MLPNN), and clustering based on Self-Organizing Map (SOM clustering). In the formulation of CDIA, these techniques are modified, improved and organized into five components and three workflows, to satisfy the demands of the conceptual design of indoor arenas.