Fuzzy C-means Research Articles

Classification of Chest Disease Detection Using X-Ray Images through the Implantation of Efficientnetv2

Abstract—Chest diseases encompass a broad spectrum of conditions that can severely affect both respiratory and cardiovascular health. These diseases may arise from various factors, including infections, environmental influences, genetic predispositions, and lifestyle choices. To address these challenges, this project proposes a system for chest disease detection using X-ray images, employing the EfficientNet V2 model.The methodology begins with preprocessing steps, where an Adaptive Median Filter is used to enhance the quality of X-ray images by reducing noise while preserving important features for analysis. After preprocessing, Fuzzy C-Means (FCM) Image Segmentation is applied to precisely identify key regions within the images, improving the robustness of the classification model. EfficientNet V2 is then employed to extract features and classify the images, offering improved accuracy compared to traditional methods.Additionally, Local Binary Patterns (LBP) are incorporated for feature extraction, further enriching the input data for the model.This proposed framework aims to automate the diagnostic process, reducing the reliance on manual interpretation by radiologists. Experimental results demonstrate the effectiveness of this approach in accurately identifying and classifying chest diseases, thereby contributing to faster and more efficient patient care. The entire project is implemented using Python. Keywords—Fuzzy C-Means (FCM),Local Binary Patterns (LBP), Convolutional Neural Network, Lung Disease.

Associations of motor and neuropsychiatric symptoms with comorbidities in prodromal Parkinson's disease.

To investigate the associations between comorbidities and multimorbidity patterns with motor and neuropsychiatric symptoms in patients with Parkinson's disease (PD) in prodromal PD. Multimorbidity is defined as the coexistence of two or more long-term conditions (LTCs) (also known as multiple comorbidities). A total of 921 participants without PD were included in the Parkinson's Progression Markers Initiative (PPMI) database and were categorized according to the LTC count. Participants were evaluated on motor and psychiatric symptoms. Pearson correlation to examine relationship of comorbidities and target symptoms. The baseline population was analyzed using Multiple linear regression model, while mixed effects model was utilized for longitudinal analysis. Fuzzy C-means clustering analysis was conducted to identify comorbidity patterns, followed by multiple linear regression for further analysis. At baseline, a higher LTC count was significantly correlated with more severe motor (MDS-UPDRS I, II, ADL, all P < 0.05) and neuropsychiatric symptoms (QUIP, P < 0.001). Three multimorbidity patterns were identified. Among them, the cardiometabolic multimorbidity pattern (CAR) had the most significant correlation with the aforementioned symptoms. Our longitudinal analysis indicated that an increase in the LTC count was associated with the exacerbation of motor and neuropsychiatric symptoms. Comorbidities were cross-sectionally and longitudinally associated with the motor and neuropsychiatric symptoms of patients with prodromal PD. Among the three multimorbidity patterns, CAR posed the highest threat to the risk of more severe motor and neuropsychiatric symptoms.

PMWFCM: A Possibility based MultiKernel Weighted Fuzzy Clustering Algorithm for classification of driving behaviors

Fuzzy clustering algorithms are widely applied in the field of traffic driving, aiding in the classification of driving behaviors from massive traffic data and enhancing traffic safety levels. However, classical Fuzzy C-Means (FCM) algorithms are sensitive to noise during the clustering process, leading to suboptimal performance when dealing with traffic datasets with lower accuracy. Moreover, single-kernel clustering algorithms are greatly influenced by kernel function selection. To address these issues, this paper proposes a Possibility Weighted Multi-Kernel Fuzzy Clustering Algorithm (PMWFCM). By integrating possibility-based fuzzy clustering with FCM and introducing a multi-kernel weighting mechanism, PMWFCM effectively reduces FCM’s sensitivity to outliers while resolving issues of clustering consistency in Possibilistic C-Means (PCM) algorithms, overcoming the challenges associated with kernel function selection. Validation on three different types of datasets demonstrates that the PMWFCM algorithm performs exceptionally well in terms of average accuracy, normalized information, average time, robustness, and convergence. When applied to the evaluation of driving behaviors in traffic datasets. Therefore, the improved FCM algorithm proposed in this paper can accurately and comprehensively reflect changes in traffic data, providing a solid theoretical foundation for identifying and assessing major risk types among passenger drivers.

Big data processing using hybrid Gaussian mixture model with salp swarm algorithm

The traditional methods used in big data, like cluster creation and query-based data extraction, fail to yield accurate results on massive networks. To address such issues, the proposed approach involves using the Hadoop Distributed File System (HDFS) for data processing, the map-reduce programming paradigm for data processing, and query optimization techniques to quickly and effectively extract accurate outcomes from a variety of options with a high processing capacity. The methodology proposed in this work makes use of Gaussian Mixture Model (GMM) for data clustering and the Salp Swarm Algorithm (SSA) for optimization. The security of preprocessed data stored on networked clusters with interconnections has been ensured by SHA algorithms. Finally, incorporating into consideration the important parameters, evaluation findings for the experimental performance of the model in the indicated methodology are produced. For this work, the estimated range of input file sizes is 60–100 MB. The processing of 100 MB of input files yielded an accuracy of 96% and results for specificity and sensitivity of 90% and 93%, respectively. The outcomes have been compared with well-known methods like fuzzy C-means and K-means approaches, and the results show that the proposed method effectively distributes accurate data processing to cluster nodes with low latency. Moreover, it uses the least amount of memory resources possible when operating on functional CPUs. As a result, the proposed approach outperforms existing techniques.

Two Fuzzy Clustering Algorithms Based on ARMA Model

This study proposes two fuzzy clustering algorithms based on autoregressive moving average (ARMA) model for series data. The first, referred to as Tsallis entropy-regularized fuzzy c-ARMA model (TFCARMA), is created from k-ARMA, a conventional hard clustering algorithm for series data. TFCARMA is motivated by the relationship between the two clustering algorithms for vectorial data: k-means and Tsallis entropy-regularized fuzzy c-means. The second, referred to as q-divergence-based fuzzy c-ARMA model (QFCARMA), is created from ARMA mixtures, a conventional probabilistic clustering algorithm for series data. QFCARMA is motivated by the relationship between the two clustering algorithms for vectorial data: Gaussian mixture model and q-divergence-based fuzzy c-means. Based on numerical experiments using an artificial dataset, we observed the effects of fuzzification parameters in the proposed algorithms and relationship between the proposed and conventional algorithms. Moreover, numerical experiments using seven real datasets compared the clustering accuracy among the proposed and conventional algorithms.

Predicting the S-ovalbumin content and determining the freshness of chicken eggs via transmittance spectroscopy

The quality and freshness of eggs deteriorate during storage, impacting their role as a crucial dietary and industrial component. This study utilizes transmittance spectroscopy (200–1000 nm) to predict the S-ovalbumin content and to identify optimal wavelengths for assessing egg freshness levels. Three hundred fresh eggs were evaluated over 21 days of storage at ambient (25 ± 2 °C, 45 ± 5% R.H.) and refrigerated (5 ± 0.5 °C, 75 ± 5% R.H.) conditions by measuring quality indicators (Haugh unit, air cell height, yolk index, S-ovalbumin, and albumin/yolk pH). The samples were divided into several clusters using the fuzzy C-means algorithm based on these parameters. The changes in the S-ovalbumin concentration increased significantly (p < 0.01) across the storage conditions. Employing the first three principal components of preprocessed transmittance spectra as inputs to artificial neural networks enabled the discrimination of fresh from old eggs with 92% accuracy and the prediction of S-ovalbumin content with R = 0.93 and RMSE = 0.09. The light transmittance at wavelengths of 228 and 280 nm, as determined via the RELIEF algorithm, was used to classify the eggs into fresh/old (74.43% accuracy) and fresh/semi-fresh/old (72.85% accuracy) groups, demonstrating the efficacy of spectroscopy for non-destructive egg quality evaluation.

Comparison of Clustering Algorithms: Fuzzy C-Means, K-Means, and DBSCAN for House Classification Based on Specifications and Price

This study aims to compare the performance of three clustering algorithms, namely Fuzzy C-Means, K-Means, and DBSCAN, in grouping houses based on their specifications and prices. The data used includes features such as price, building area, land area, number of bedrooms, number of bathrooms, and availability of garages. The performance of these algorithms was evaluated using Silhouette Score and Davies-Bouldin Score to determine the quality of cluster separation. The results indicate that K-Means achieved the best performance with the highest Silhouette Score of 0.7702 for two clusters, followed by Fuzzy C-Means, which excelled in handling overlapping clusters. DBSCAN, while effective in detecting outliers, showed suboptimal performance for this housing dataset. These findings suggest that K-Means is the most suitable clustering method for housing data, while Fuzzy C-Means and DBSCAN can serve as alternatives depending on the data characteristics. This research is expected to assist in making the house searching and classification process more efficient and provide additional insights for developers in shaping housing market strategies.

Cluster analysis via projection onto convex sets

This paper proposes a data clustering algorithm that is inspired by the prominent convergence property of the Projection onto Convex Sets (POCS) method, termed the POCS-based clustering algorithm. For disjoint convex sets, the form of simultaneous projections of the POCS method can result in a minimum mean square error solution. Relying on this important property, the proposed POCS-based clustering algorithm treats each data point as a convex set and simultaneously projects the cluster prototypes onto respective member data points, the projections are convexly combined via adaptive weight values in order to minimize a predefined objective function for data clustering purposes. The performance of the proposed POCS-based clustering algorithm has been verified through a large scale of experiments and data sets. The experimental results have shown that the proposed POCS-based algorithm is competitive in terms of both effectiveness and efficiency against some of the prevailing clustering approaches such as the K-Means/K-Means+⁣+ and Fuzzy C-Means (FCM) algorithms. Based on extensive comparisons and analyses, we can confirm the validity of the proposed POCS-based clustering algorithm for practical purposes.

Hyperspectral crop image classification via ensemble of classification model with optimal training

Agriculture is a significant source of income, and categorizing the crop has turned into vital factor that aids more in the crop production sector. Traditionally, crop development stage determination is done manually by eye inspection. However, producing high-quality crop type maps using modern approaches remains difficult. In this paper, the hyperspectral crop image classification model is proposed that includes four stages, they are (a) preprocessing, (b) segmentation, (c) feature extraction and (d) classification. In the preprocessing step, the hyperspectral image is provided as input, where the filtering process will carried out using median filtering. The filtered image is then used as the segmentation’s input. The image is segmented in the segmentation step using the enhanced entropy-based fuzzy c-means technique. Subsequently, spectral spatial features and vegetation index-based features are derived from segmented images. The final step is the classification, where the ensemble of classification model will be used that includes models like Convolutional Neural Networks (CNN), Deep Maxout (DMO), Recurrent Neural Networks (RNN), and Bidirectional Gated Recurrent Unit (Bi-GRU), respectively. The proposed Self Improved Tasmanian devil Optimization (SI-TDO) approach has optimally adjusted the Bi-GRU model’s training weights to enhance ensemble classification performance. Finally, the effectiveness of the proposed SI-TDO method compared to the traditional algorithm is examined for several metrics. The SI-TDO obtained the greatest accuracy of 94.68% in training rate 80, while other existing models have the lowest ratings.

Sonar image segmentation using a multi-spatial information constraint fuzzy C-means clustering algorithm based on KL divergence

Sonar image segmentation using a multi-spatial information constraint fuzzy C-means clustering algorithm based on KL divergence

An on-machine tool wear area identification method based on image augmentation and advanced segmentation

In industrial manufacturing, tool wear monitoring (TWM) is essential for ensuring high-quality machining, operational efficiency, cost-effectiveness, and safety. However, due to the complexities of on-machine imaging and the constraints of direct measurement techniques, TWM methods face challenges such as unclear boundaries, class imbalance between wear and unworn area, and image diversity. This paper proposes a novel two-step approach for identifying tool wear areas. Firstly, DeepLabV3+ with Focal Loss is utilized to identify the Region of Interest (ROI) of the tool. Secondly, the method employs Intuitionistic Fuzzy C-Means Clustering (IFCM) for detailed segmentation of the wear area. This integration effectively addresses challenges arising from uneven illumination that blur image boundaries and the class imbalance between images with tool wear and those without. To enhance image diversity and quality, we utilize Denoising Diffusion Probabilistic Models (DDPM) for image augmentation, significantly enriching the training dataset. The proposed approach achieves a Mean Pixel Accuracy (MPA) of 95.32% and a Mean Intersection over Union (MIoU) of 93.67%, which marks a substantial improvement over existing TWM models. This progress not only provides a more reliable and efficient tool wear monitoring solution but also sets a new standard for precision in industrial machining processes.

Retinal blood vessel segmentation using density-based fuzzy C-means clustering and vessel neighborhood connected component

Retinal blood vessel segmentation is a crucial process in medical image analysis. However, it is often challenging due to the variation in color, shape, intensity, size, and contrast of blood vessels. Most of the relevant studies concentrate on algorithms based on supervised learning and few on deep learning. However, due to the several challenges in retinal image acquisition, these algorithms cannot deliver the highest possible level of accuracy. Therefore, this paper implements retinal blood vessel segmentation and classification (RBVSC) methods using density-based fuzzy C-means clustering and vessel neighborhood-connected components, hereafter denoted as DBFCM-VNCC. Initially, the given retinal images are preprocessed using the contrast enhancement method that involves Histogram Equalization with Variable Enhancement Degree (HEVED), selecting the appropriate color channel, optic disc elimination, and using a Gaussian filter, which removes the noise or artifacts from the retinal images. Then, a fully fuzzy C-means clustering is used for coarse segmentation of vessel lesions, which can detect the affected blood vessel features quite efficiently. Finally, structure-based algorithms based on vessel neighborhood-connected components based on mathematical dilation operators and local thicknesses are used to obtain accurate skeletonization and segmentation of the retinal vessels. The algorithm was assessed using three open-access retinal image databases: DRIVE, CHASE_DB1, and HRF, where it achieved mean sensitivity, specificity, accuracy, area overlap measure, and error rate scores of 98.16%, 98.74%, 97.68%, and 4.54%; 98.25%, 98.81%, 97.68%, 97.86%, and 2.14%; 98.22%, 98.78%, 97.56%, 97.40%, and 2.60% for segmenting the retinal vessel. This demonstrates the effectiveness of the proposed DBFCM-VNCC techniques.

Research on stacking ensemble method for day-ahead ultra-short-term prediction of photovoltaic power

The intermittency and uncertainty of photovoltaic power output make it difficult to match real-time electricity demand, and accurate prediction of photovoltaic power is the basis for improving photovoltaic energy consumption. This article proposes a new stacking ensemble forecast model based on the time series generative adversarial network (TimeGAN-SEFM), to improve the accuracy of ultra-short-term forecasting of photovoltaic power. Firstly, a dual layer feature weighted clustering weighted fuzzy C-means clustering algorithm is proposed to accurately partition the original dataset according to weather types. Secondly, in the proposed TimeGAN-SEFM, a Particle Swarm Optimization based BP neural network is used as a meta learner to integrate three types of base learners, thereby improving the structural diversity of the model. The dataset for each type of weather is proportionally divided into multiple sub-datasets for training base learners. In response to the limitation that the amount of data used for training a single base learner is relatively small after multiple splitting of the original data, a TimeGAN augmentation network is introduced to perform multi-dimensional data augmentation on the sub-training sets, which improves the size and diversity of the dataset used for training base learners, thereby improving the predictive performance of the model. Finally, the predictive performance of the proposed TimeGAN-SEFM was evaluated using measurement data from a 20 MW photovoltaic power station in North China. The results show that TimeGAN-SEFM effectively improves the accuracy of photovoltaic prediction and helps promote real-time matching of photovoltaic energy and power demand, thereby improving the utilization of photovoltaic energy.

Hybrid regret-based p-robust and distributionally robust optimization models for electric vehicle charging station network design

The rise of electric vehicles (EVs) in recent years has underscored the need for well-established Electric Vehicle Charging Station (EVCS) infrastructure. The lack of a reliable network of EVCS is a significant obstacle to the widespread adoption of EVs. Also, one often overlooked aspect of electric vehicles is the importance of using renewable energy sources to recharge their batteries. Without sustainable energy, EVs could produce more greenhouse gas emissions than traditional vehicles. Therefore, it seems essential to design networks of EVCSs that depend on renewable sources to meet their energy needs and ensure convenient access for customers while reducing environmental impact. This study proposes a two-stage stochastic optimization model to determine the optimal locations and capacities for EVCSs and supporting sustainable energy plants. The model considers uncertain factors such as vehicle flows and renewable energy supply. The model combines a hybrid adjusted p-robust and min–max regret approach to address uncertainties to ensure robustness in the network design process. Two approaches including Sample Average Approximation (SAA) and Fuzzy C-means Clustering are employed to solve the model efficiently. Additionally, the Distributionally Robust Optimization (DRO) approach accounts for uncertainty in the probability distribution function of scenarios. The results of these solution approaches are validated through parameter realization, which involves simulating uncertain parameters and verifying the robustness of the solutions obtained from each approach. This validation process is done by comparing the obtained results with the optimal values based on two criteria: the sum of deviation from optimality and the maximum deviation from optimality. By adopting these solution methods, the proposed methodology contributes to developing efficient and sustainable EV infrastructure. This methodology gives investors and city officials robust decision support for establishing EVCS.

Handling Noise Data with PCA Method and Optimization Using Hybrid Fuzzy C-Means and Genetic Algorithm

The significance of machine learning (ML) and data mining techniques particularly clustering is examined in this research, in managing large data sets for customer segmentation in the retail sector. The research emphasizes the challenges posed by data noise and proposes a solution using Principal Component Analysis (PCA) to improve accuracy. This study introduces a hybrid approach that combines Fuzzy C-Means (FCM) with genetic algorithms for optimization in customer segmentation, and suggests further research on the optimal number of clusters and data noise elimination. By addressing data noise, the proposed PCA-based method achieved a higher accuracy rate of 98% compared to 93% without PCA. This finding underscores the effectiveness of PCA in noise reduction, improving clustering accuracy. This research contributes to the advancement of customer-focused business strategies through better data analysis and interpretation. The proposed approach has potential applications in areas including data analysis, pattern recognition, and image processing, highlighting its relevance in the contemporary business environment.

Watershed Grouping Based on Suspended Sediment Yield Using Morphological Criteria: A Comparison of Hierarchical and Nonhierarchical Techniques

ABSTRACTRegional sediment analysis methods have traditionally been used to reliably estimate sediment yield in ungauged watersheds with adequate hydrometric stations. Watershed‐grouping techniques are a suitable approach for estimating sediment load, which can be used to assess erosion processes, and surface water quality, and plan soil and water conservation measures. Additionally, the relationship between sediment yield and watershed characteristics can be used to identify different stages of erosion evolution and prioritize watersheds for conservation efforts. To this end, this study aimed to assess the applicability of various analytical hierarchy process (AHP) methods, including Ward's, Single linkage, and β‐flexible methods, as well as the nonhierarchical fuzzy C‐means (FCM) method in the Gorganroud and Qareh‐Sou watersheds of northern Iran based on hydrological, geological, climatic, physiographical, and land‐use characteristics. Furthermore, the study evaluated clustering methods to determine the optimal number of subwatershed clusters. The AHP methods identified two, three, or four clusters, while the FCM method identified three clusters as the best number of clusters. The results indicated that the FCM method had a well‐regionalized estimation in the study area with the maximum amount of validation indices (The highest Dunn coefficient 0.31–0.65, the lowest TESS 16.62–36.8, and acceptable Pseudo‐F coefficient 8.1–10.4). Notably, the fuzzy clustering methods associated each input sample with two or more clusters, which reduced the uncertainty of sediment yield estimation. The recommended clustering method can be used for regional sediment analysis in ungauged watersheds and evaluated against other hydrological variables in the study area.

ANALYSIS OF FUZZY C-MEANS IN PERSONALITY CLUSTERING BASED ON THE OCEAN MODEL

Personality is the pattern of an individual's behavior in daily life, reflected in their thoughts, feelings, and actions. The Big Five Personality Traits Model, known as OCEAN, helps to understand the complexity of human personality through five main traits. The identification and classification of personality, particularly among students, impacts academic performance, personal development, anxiety levels, and risky behaviors. Collaboration between educators, mental health professionals, and career advisors is crucial to creating an educational environment that supports students' holistic development. The Fuzzy C-Means (FCM) method is used to identify students' personalities with adequate accuracy. This study adopts the OCEAN model with FCM to efficiently identify and classify students' personalities. Data were obtained from 142 respondents, resulting in 27% of respondents being classified in cluster 1, 21% in cluster 2, 18% in cluster 3, 16% in cluster 4, and 18% in cluster 5. This study has important implications for students, educators, and educational institutions to understand that learning patterns, social interactions, and decision-making processes can be influenced by an individual's personality.

Crude Oil Price Forecasting Using Hybrid Heuristic Model and Fuzzy C-Means on Type 2 Fuzzy Time Series

Forecasting world crude oil prices needs to be done because it has an essential role in Indonesia’s economy, so an accurate and efficient forecasting approach is needed. This research combines heuristic and Fuzzy C-Means (FCM) models on Type 2 Fuzzy Time Series (T2FTS) to forecast crude oil prices. T2FTS, an extension of Fuzzy Time Series (FTS) by adding observations, is used to enrich the fuzzy relationship of the Type 1 model and can improve forecasting performance. FCM is used to determine unequal interval lengths and heuristic models to optimize fuzzy relations by identifying crude oil price movements using up and down trends. The data used in this study is the price of Brent crude oil as one of the benchmarks for world crude oil prices from 1 January 2023 – 31 May 2024. Mean Absolute Percentage Error (MAPE) measures accuracy in assessing forecasting results. The results showed that the combination of heuristic and FCM models in T2FTS gave accurate results, as evidenced by the MAPE value obtained, which was 1.50%, so it fell into the excellent category.

Modal Parameter Identification of Electric Spindles Based on Covariance-Driven Stochastic Subspace

Electric spindles are a critical component of numerically controlled machine tools that directly affect machining precision and efficiency. The accurate identification of the modal parameters of an electric spindle is essential for optimizing design, enhancing dynamic performance, and facilitating fault diagnosis. This study proposes a covariance-driven stochastic subspace identification (SSI-cov) method integrated with a simulated annealing (SA) strategy and fuzzy C-means (FCM) clustering algorithm to achieve the automated identification of modal parameters for electric spindles. Using both finite element simulations and experimental tests conducted at 22 °C, the first five natural frequencies of the electric spindle under free, constrained, and dynamic conditions were extracted. The experimental results demonstrated experiment errors of 0.17% to 0.33%, 1.05% to 3.27%, and 1.29% to 3.31% for the free, constrained, and dynamic states, respectively. Compared to the traditional SSI-cov method, the proposed SA-FCM method improved accuracy by 12.05% to 27.32% in the free state, 17.45% to 47.83% in the constrained state, and 25.45% to 49.12% in the dynamic state. The frequency identification errors were reduced to a range of 2.25 Hz to 20.81 Hz, significantly decreasing errors in higher-order modes and demonstrating the robustness of the algorithm. The proposed method required no manual intervention, and it could be utilized to accurately analyze the modal parameters of electric spindles under free, constrained, and dynamic conditions, providing a precise and reliable solution for the modal analysis of electric spindles in various dynamic states.

An effective Key Frame Extraction technique based on Feature Fusion and Fuzzy-C means clustering with Artificial Hummingbird

Key frame extraction is very important in video summarization and content-based video analysis to address the problem of data redundancy in a video. Key frame extraction enables quick navigation and expert video arrangement in many applications. The visually impaired can benefit from the use of key frame extraction for rapid object recognition and tracking. Most key frame extraction techniques consider only a single visual feature instead of multiple features or full pictorial information of the video. This study proposes a key frame extraction method from a video that (i) first removes insignificant frames by pre-processing, (ii) second, four visual and structural feature differences among the consecutive frames are extracted and aggregated to identify informative frames, (iii) third, to cluster the obtained frames, a hybrid FCM-AHA method is proposed by combining Fuzzy C-means(FCM) with artificial hummingbird optimization algorithm (AHA) to circumvent the local minima trapping problem of FCM, and finally, from each cluster, the two frames having greatest Euclidean distance from all the other frames within a cluster is selected as key frames to remove redundant frames. Experimental results on the Open video and YouTube datasets show that the suggested method outperforms state-of-the-art methods both in terms of subjective qualitative analysis and objective quantitative evaluation, e.g., Precision, Recall, and F-score. Further, results are also taken on real video to demonstrate its applicability in real-life applications.