A novel self-organizing map (SOM) neural network for discrete groups of data clustering

Unsupervised Clustering-Assisted Method for Consensual Quantitative Analysis of Methanol-Gasoline Blends by Raman Spectroscopy

Methanol-gasoline blends have emerged as a promising and environmentally friendly bio-fuel option, garnering widespread attention and promotion globally. The methanol content within these blends significantly influences their quality and combustion performance. This study explores the qualitative and qualitative analysis of methanol-gasoline blends using Raman spectroscopy coupled with machine learning methods. Experimentally, methanol-gasoline blends with varying methanol concentrations were artificially configured, commencing with initial market samples. For qualitative analysis, the partial least squares discriminant analysis (PLS-DA) model was employed to classify the categories of blends, demonstrating high prediction performance with an accuracy of nearly 100% classification. For the quantitative analysis, a consensus model was proposed to accurately predict the methanol content. It integrates member models developed on clustered variables, using the unsupervised clustering method of the self-organizing mapping neural network (SOM) to accomplish the regression prediction. The performance of this consensus model was systemically compared to that of the PLS model and uninformative variable elimination (UVE)-PLS model. Results revealed that the unsupervised consensus model outperformed other models in predicting the methanol content across various types of methanol gasoline blends. The correlation coefficients for prediction sets consistently exceeded 0.98. Consequently, Raman spectroscopy emerges as a suitable choice for both qualitative and quantitative analysis of methanol-gasoline blend quality. This study anticipates an increasing role for Raman spectroscopy in analysis of fuel composition.

Generic and accurate prediction of retention times in liquid chromatography by post-projection calibration

Retention time predictions from molecule structures in liquid chromatography (LC) are increasingly used in MS-based targeted and untargeted analyses, providing supplementary evidence for molecule annotation and reducing experimental measurements. Nevertheless, different LC setups (e.g., differences in gradient, column, and/or mobile phase) give rise to many prediction models that can only accurately predict retention times for a specific chromatographic method (CM). Here, a generic and accurate method is present to predict retention times across different CMs, by introducing the concept of post-projection calibration. This concept builds on the direct projections of retention times between different CMs and uses 35 external calibrants to eliminate the impact of LC setups on projection accuracy. Results showed that post-projection calibration consistently achieved a median projection error below 3.2% of the elution time. The ranking results of putative candidates reached similar levels among different CMs. This work opens up broad possibilities for coordinating retention times between different laboratories and developing extensive retention databases.

A supervised method to enhance distance-based neural network clustering performance by discovering perfect representative neurons

Distance-based neural network clustering requires the intrinsic assumption that a particular neuron in the network represents a cluster centroid. However, not all these neurons can perfectly represent the training data; these neurons can only represent part of the training samples. This paper proposes an effective training data splitting method (TDSM) to find perfect representative neurons and improve the clustering results in a distance-based neutral network without changing the original network’s internal algorithm or the training data quality. The method allows a network with N neurons to be enlarged to a new network with $$m\times N$$ m × N neurons. These neurons represent m subnetworks, and each subnetwork perfectly represents a part of the training set, where the clustering qualification indicators (the purity, normalized mutual information, and adjusted rand index measures) all equal 1. The results are statistically validated with a t test, and we demonstrate that the TDSM performs better than the original clustering paradigm on some real datasets.

The potential ecological risk assessment of soil heavy metals using self-organizing map

Heavy metal pollution control zoning aiming at the health maintenance of watershed soil ecosystem has become an important means of soil environmental protection and governance. Based on the self-organizing map (SOM), this study classifies the data sets of eight heavy metals such as Co, Cd, Zn, Cr, Cu, Pb, Ni, and Tl in 354 samples, calculates the potential ecological risk value of soil heavy metals in combination with the potential Hakansom Risk index (HRI), and uses the geographic information system (GIS) for visualization. In the research results, SOM has divided five soil ecological risk categories. The highest average ecological risk value of 85.95 is found in cluster IV, which is clustered and distributed in urban development areas in the upper reaches of the river. The average ecological risk values of cluster I and cluster V are relatively close at 79.64 and 79.19, respectively. Cluster I and cluster V are distributed in the north of the river in a linear and cluster manner, respectively, and are located on a concave bank with a relatively gentle slope. The average ecological risk of soil pollution in cluster II is 77.59, which is linearly distributed on both banks of the river. The ecological risk of soil pollution in cluster III is the lowest (74.39), mainly scattered in the south of rivers with less human activities. The study further identified the environmental factors that affect the soil ecological risk value in different cluster units and put forward the classified and differentiated management and control strategies for different cluster units. The research shows that SOM can cluster the data sets of heavy metals with high sensitivity and low threshold through competitive learning to effectively provide the distribution information of abnormal soil ecological risk areas. This information is helpful for urban environmental management departments and planning departments to take targeted management and recovery measures to avoid the health risks related to soil heavy metal pollution.

Weather Conditions and COVID-19 Cases: Insights from the GCC Countries

The prediction of new COVID-19 cases is crucial for decision makers in many countries. Researchers are continually proposing new models to forecast the future tendencies of this pandemic, among which long short-term memory (LSTM) artificial neural networks have exhibited relative superiority compared to other forecasting techniques. Moreover, the correlation between the spread of COVID-19 and exogenous factors, specifically weather features, has been explored to improve forecasting models. However, contradictory results have been reported regarding the incorporation of weather features into COVID-19 forecasting models. Therefore, this study compares uni-variate with bi- and multi-variate LSTM forecasting models for predicting COVID-19 cases, among which the latter models consider weather features. LSTM models were used to forecast COVID-19 cases in the six Gulf Cooperation Council countries. The root mean square error (RMSE) and coefficient of determination (R2) were employed to measure the accuracy of the LSTM forecasting models. Despite similar weather conditions, the weather features that exhibited the strongest correlation with COVID-19 cases differed among the six countries. Moreover, according to the statistical comparisons that were conducted, the improvements gained by including weather features were insignificant in terms of the RMSE values and marginally significant in terms of the R2 values. Consequently, it is concluded that the uni-variate LSTM models were as good as the best bi- and multi-variate LSTM models; therefore, weather features need not be included. Furthermore, we could not identify a single weather feature that can consistently improve the forecasting accuracy.

A faster dynamic convergency approach for self-organizing maps

This paper proposes a novel variable learning rate to address two main challenges of the conventional Self-Organizing Maps (SOM) termed VLRSOM: high accuracy with fast convergence and low topological error. We empirically showed that the proposed method exhibits faster convergence behavior. It is also more robust in topology preservation as it maintains an optimal topology until the end of the maximum iterations. Since the learning rate adaption and the misadjustment parameter depends on the calculated error, the VLRSOM will avoid the undesired results by exploiting the error response during the weight updation. Then the learning rate is updated adaptively after the random initialization at the beginning of the training process. Experimental results show that it eliminates the tradeoff between the rate of convergence and accuracy and maintains the data's topological relationship. Extensive experiments were conducted on different types of datasets to evaluate the performance of the proposed method. First, we experimented with synthetic data and handwritten digits. For each data set, two experiments with a different number of iterations (200 and 500) were performed to test the stability of the network. The proposed method was further evaluated using four benchmark data sets. These datasets include Balance, Wisconsin Breast, Dermatology, and Ionosphere. In addition, a comprehensive comparative analysis was performed between the proposed method and three other SOM techniques: conventional SOM, parameter-less self-organizing map (PLSOM2), and RA-SOM in terms of accuracy, quantization error (QE), and topology error (TE). The results indicated the proposed approach produced superior results to the other three methods.

Oil Family Typing Using a Hybrid Model of Self-Organizing Maps and Artificial Neural Networks

Identifying the number of oil families in petroleum basins provides practical and valuable information in petroleum geochemistry studies from exploration to development. Oil family grouping helps us track migration pathways, identify the number of active source rock(s), and examine the reservoir continuity. To date, almost in all oil family typing studies, common statistical methods such as principal component analysis (PCA) and hierarchical clustering analysis (HCA) have been used. However, there is no publication regarding using artificial neural networks (ANNs) for examining the oil families in petroleum basins. Hence, oil family typing requires novel and not overused and common techniques. This paper is the first report of oil family typing using ANNs as robust computational methods. To this end, a self-organization map (SOM) neural network associated with three clustering validity indexes was employed on oil samples belonging to the Iranian part of the Persian Gulf oilfields. For the SOM network, at first, 10 default clusters were selected. Afterward, three effective clustering validity coefficients, namely, Calinski-Harabasz (CH), Silhouette (SH), and Davies-Bouldin (DB), were studied to find the optimum number of clusters. Accordingly, among 10 default clusters, the maximum CH (62) and SH (0.58) were acquired for 4 clusters. Similarly, the lowest DB (0.8) was obtained for four clusters. Thus, all three validation coefficients introduced four clusters as the optimum number of clusters or oil families. According to the geochemical parameters, it can be deduced that the corresponding source rocks of four oil families have been deposited in a marine carbonate depositional environment under dysoxic-anoxic conditions. However, oil families show some differences based on geochemical data. The number of oil families identified in the present report is consistent with those previously reported by other researchers in the same study area. However, the techniques used in the present paper, which have not been implemented so far, can be introduced as more straightforward for clustering purposes in oil family typing than those of common and overused methods of PCA and HCA.

Evaluating the effect of quran memorizing on the event-related potential features by using graphs created from the neural gas networks

Background:

Quran memorizing causes a state of trance, which its result is the changes in the amplitude and time of P300 and N200 components in the event related potential (ERP) signal. Nevertheless, a limited number of studies that have examined the effects of Quran memorizing on brain signals to enhance relaxation and attention, and improve the lives of patients with autism and stroke, generally have not presented any analysis based on comparing structural differences relevant to features extracted from ERP signal obtained from the two groups of Quran memorizer and nonmemorizer by using the hybrid of graph theory and competitive networks.

Methods:

In this study, we investigated structural differences relevant to the graph obtained from the weight of neural gas (NG) and growing NG (GNG) networks trained by features extracted from the ERP signal recorded from two groups during the PRM test. In this analysis, we actually estimated the ERP signal by averaging the brain background data in the recovery phase. Then, we extracted six features related to the power and the complexity of these signals and selected optimal channels in each of the features by using the t test analysis. Then, these features extracted from the optimal channels are applied for developing the NG and GNG networks. Finally, we evaluated different parameters calculated from graphs, in which their connection matrix was obtained from the weight matrix of the networks.

Results:

The outcomes of this analysis show that increasing the power of low frequency components and the power ratio of low frequency components to high frequency components in the memorizers, which represents patience, concentration, and relaxation, is more than that of the nonmemorizers. These outcomes also show that the optimal channels in different features, which were often in frontal, peritoneal, and occipital regions, had a significant difference (P < 0.05). It is remarkable that two parameters of the graphs established based on two competitive networks, i.e. average path length and the average of the weights in the memorizers, were larger than the nonmemorizers, which means more data scattering in this group.

Conclusion:

This condition in the mentioned graphs suggests that the Quran memorizing causes a significant change in ERP signals, so that its features have usually more scattering.

Anomaly Detection in Automotive Industry Using Clustering Methods—A Case Study

In automotive industries, pricing anomalies may occur for components of different products, despite their similar physical characteristics, which raises the total production cost of the company. However, detecting such discrepancies is often neglected since it is necessary to find the problems considering the observation of thousands of pieces, which often present inconsistencies when specified by the product engineering team. In this investigation, we propose a solution for a real case study. We use as strategy a set of clustering algorithms to group components by similarity: K-Means, K-Medoids, Fuzzy C-Means (FCM), Hierarchical, Density-Based Spatial Clustering of Applications with Noise (DBSCAN), Self-Organizing Maps (SOM), Particle Swarm Optimization (PSO), Genetic Algorithm (GA) and Differential Evolution (DE). We observed that the methods could automatically perform the grouping of parts considering physical characteristics present in the material master data, allowing anomaly detection and identification, which can consequently lead to cost reduction. The computational results indicate that the Hierarchical approach presented the best performance on 1 of 6 evaluation metrics and was the second place on four others indexes, considering the Borda count method. The K-Medoids win for most metrics, but it was the second best positioned due to its bad performance regarding SI-index. By the end, this proposal allowed identify mistakes in the specification and pricing of some items in the company.

Overcome chromatism of metasurface via Greedy Algorithm empowered by self-organizing map neural network

Chromatism generally exists in most metasurfaces. Because of this, the deflected angle of metasurface reflectors usually varies with frequency. This inevitably hinders wide applications of metasurfaces to broadband signal scenarios. Therefore, it is of great significance to overcome chromatism of metasurfaces. With this aim, we firstly analyze necessary conditions for achromatic metasurface deflectors (AMD) and deduce the ideal dispersions of meta-atoms. Then, we establish a Self-Organizing Map (SOM) Neural Network as a prepositive model to obtain a diversified searching map, which is then applied to Greedy Algorithm to search meta-atoms with the required dispersions. Using these meta-atoms, an AMD was designed and simulated, with a thickness about 1/15 the central wavelength. A prototype was fabricated and measured. Both the simulation and measurement show that the proposed AMD can achieve an almost constant deflected angle of 22° under normal incidence within 9.5-10.5GHz. This method may find wide applications in designing functional metasurfaces for satellite communications, mobile wireless communications and others.

Differentiating Thamnocalamus Munro from Fargesia Franchet emend. Yi (Bambusoideae, Poaceae): novel evidence from morphological and neural-network analyses

Fargesia Franchet emend. Yi is closely allied with Thamnocalamus Munro but differs in many major morphological characteristics. Based on traditional morphological characters, it is difficult to differentiate these two genera. The current study measured 19 species in these two genera to determine whether variations in 12 categories of major characters are continuous. In addition, a self-organizing map (SOM) and cluster analysis were used together to reveal whether the known species of Fargesia represent discontinuous sampling of Thamnocalamus. The results show that 46 morphological characteristics exhibited high variation at the generic and species levels. In addition, the cluster analysis showed that 32 morphological characteristics of Thamnocalamus and Fargesia were divided between two species and well separated from the outgroup. Additionally, significant differences (P < 0.01) were observed in the reproductive structures between these two genera. The unrooted dendrogram, which was based on the SOM neural network, shows the same results as the cluster analysis of morphological characteristics. These data indicate that Fargesia is not a result of discontinuous sampling of Thamnocalamus; thus, Fargesia should not be treated as a synonym for Thamnocalamus.

Technical note: advantages of the self-organizing controller for high-pressure sterilization equipments

A study of a self-organizing controller is implemented in a way that response to controlled system follows the desired given by the model. The self-organizing controller has proven to be a valuable tool in sterilization equipment in order to verify the capacity of the response to any change in the pressure or temperature. Basically, this type of controller is based on the Self-Organizing Map (SOM) that is a neural network algorithm of unsupervised learning. The new ideas include clustering visualization, interactive training and one-dimension arrays.

Probability-based text clustering algorithm by alternately repeating two operations

Owing to the rapid advance of internet technology, users have to face to a large amount of raw data from the World Wide Web every day, most of which is displayed in text format. This situation brings a great demand for efficient text analysis techniques by internet users. Since clustering is unsupervised and requires no prior knowledge, it is extensively adopted to help analyse textual data. Unfortunately, as far as I know, almost all the clustering algorithms proposed so far fail to deal with large-scale text collection. For precisely classifying large-scale text collection, a novel probability based text clustering algorithm by alternately repeating two operations (abbreviated as PTCART) is proposed in this paper. This algorithm just repeats two operations of (a) feature set construction and (b) text partition until the optimal partition is reached. Its convergent capacity is also validated. Experiments results demonstrate that, compared with several popular text clustering algorithms, PTCART has excellent performance.