Random Walk Algorithm Research Articles

Predicting Erosion Damage in a Centrifugal Fan

Erosion damage can occur in fans and blowers during industrial processes, cooling, and mine ventilation. This study focuses on investigating erosion caused by particulate air flows in a centrifugal fan with forward-inclined blades. This type of fan is particularly vulnerable to erosion due to its radial flow component and flow recirculation. The flow field was solved separately, and the data transferred to the particle trajectory and erosion code. This in-house code implements the Lagrangian approach and the random walk algorithm, including statistical descriptions of particle sizes, release positions, and restitution factors. The study involved two types of dust particles, with a concentration between 100 and 500 μg/m3: The first type is the Saharan (North Africa) dust, which has a finer size between 0.1 and 100 microns. The second type is the Coarse Arizona Road Dust, also known as AC-coarse dust, which has a larger size ranging from 1 to 200 microns. The complex flow conditions within the impeller and scroll, as well as the concentration and size distribution of particles, are shown to affect the paths, impact conditions, and erosion patterns. The outer wall of the scroll is most heavily eroded due to high-impact velocities by particles exiting the impeller. Erosion is more pronounced on the pressure side of the full blades compared to the splitters and casing plate. The large non-uniformities of erosion patterns indicate a strong dependence with the blade position around the scroll. Therefore, the computed eroded mass is cumulated and averaged for all the surfaces of components. These results provide useful insights for monitoring erosion wear in centrifugal fans and selecting appropriate coatings to extend the lifespan.

TEA+ : A Novel Temporal Graph Random Walk Engine with Hybrid Storage Architecture

Many real-world networks are characterized by being temporal and dynamic, wherein the temporal information signifies the changes in connections, such as the addition or removal of links between nodes. Employing random walks on these temporal networks is a crucial technique for understanding the structural evolution of such graphs over time. However, existing state-of-the-art sampling methods are designed for traditional static graphs, and as such, they struggle to efficiently handle the dynamic aspects of temporal networks. This deficiency can be attributed to several challenges, including increased sampling complexity, extensive index space, limited programmability, and a lack of scalability. In this article, we introduce TEA+ , a robust, fast, and scalable engine for conducting random walks on temporal graphs. Central to TEA+ is an innovative hybrid sampling method that amalgamates two Monte Carlo sampling techniques. This fusion significantly diminishes space complexity while maintaining a fast sampling speed. Additionally, TEA+ integrates a range of optimizations that significantly enhance sampling efficiency. This is further supported by an effective graph updating strategy, skilled in managing dynamic graph modifications and adeptly handling the insertion and deletion of both edges and vertices. For ease of implementation, we propose a temporal-centric programming model, designed to simplify the development of various random walk algorithms on temporal graphs. To ensure optimal performance across storage constraints, TEA+ features a degree-aware hybrid storage architecture, capable of adeptly scaling in different memory environments. Experimental results showcase the prowess of TEA+ , as it attains up to three orders of magnitude speedups compared to current random walk engines on extensive temporal graphs.

An innovative damping random walk algorithm with compulsive evolution for the synthesis of heat exchanger networks

Heat exchanger network synthesis in process system engineering poses a significant challenge, primarily because of the intricacies of stream matches and the nonlinear nature of continuous variables. Numerous heuristic methods are frequently trapped in local optima because of their greedy acceptance criteria, which limits their capacity to explore diverse solutions. This aspect of enhancing algorithms through acceptance Criterion has yet to be considered in previous studies. Hence, a damping strategy using damping perfect solutions is proposed and applied to the Random Walk Algorithm with Compulsive Evolution, which has recently received considerable attention. This strategy aims to prevent trapping in local optima and improve global search capabilities, thereby reducing the risk of premature convergence to the local optima. To address the impact of different damping coefficients within the strategy on the optimization process and structural variations in the algorithm, three damping strategies were designed: a fixed damping strategy with a constant damping coefficient, a variable damping strategy featuring variable damping coefficients at different optimization stages, and an adaptive damping strategy that tracks structural changes and selectively delays modifications driven by continuous variables to preserve those driven by integer variables. The results obtained from applying these strategies to three classical industrial HEN cases are 2,890,884 $/yr, 6,650,082 $/yr, and 1,709,205 $/yr, respectively. These results outperformed the optimal outcomes documented in the literature, demonstrating the effectiveness of the new method.

Graph-based phishing detection: URLGBM model driven by machine learning

Phishing attacks are a form of social engineering that involves the transmission of deceptive communications imitating a trustworthy source to deceive users into revealing confidential information. Antiphishing systems have made significant strides in recent years, allowing internet users to secure confidential and private information against such attacks. In this study, we propose a URL graph representation based on a random walk algorithm, specifically PageRank, for weighting URL tokens. To create the graph, an imaginary walker visits the URL tokens one at a time and assigns a value to each token based on the probability of encountering the target URL during the walk. We studied different random walk (rw) variations and their effects on the URL string. The BM25 algorithm was employed to produce a sparse matrix for the classification task from the token scores obtained. Experiments conducted with logistic regression revealed that the proposed model achieved an accuracy of 98.98%, a false alarm rate of 1.72%, and a missing alarm rate of 0.302%. The model also attained a 97.17% accuracy on a benchmark dataset.

Construction of ceRNA network mediated by circRNAs screening from microarray and identification of novel biomarkers for myasthenia gravis

BackgroundRecent studies have revealed that circRNA can serve as ceRNA to participate in multiple autoimmune diseases. Our study aims to explore the key circRNA as ceRNA and biomarker for MG. Methods: We used circRNA microarray to explore differentially expressed circRNAs (DECs) from MG and compare with control. Then, we predicted the target miRNA associated with DECs and screened miRNAs by the algorithm of random walk with restart (RWR). Next, we constructed the circRNA-miRNA-mRNA ceRNA regulated network (CMMC) to identify the hub objects. Following, we detected the expression of hub-circRNAs by RT-PCR. We verify has_circ_0004183 (circFRMD4) sponging miR-145-5p regulate cells proliferation using luciferase assay and CCK-8. Results: We found that the expression level of circFRMD4 and has_circ_0035381 (circPIGB) were upregulated and has_circ_0089153(circ NUP214) had the lowest expression level in MG. Finally, we proved circFRMD4 sponging miR-145-5p regulate Jurkat cells proliferation. CircFRMD4 take part in the genesis and development of MG via circFRMD4/miR145-5p axis. Conclusions: We found that circFRMD4, circPIGB and circNUP214 can be considered as valuable potential novel biomarkers for AchR + MG. CircFRMD4 participate in the development of AchR + MG via targeting binding with miR-145-5p.

TemporalHAN: Hierarchical attention-based heterogeneous temporal network embedding

Heterogeneous temporal network embedding aims to learn each node of different types of a heterogeneous temporal network in each snapshot into a low-dimensional vector representation, which can be used for various network analysis tasks such as node classification and relationship prediction. Our work proposes a novel heterogeneous temporal graph neural network embedding framework (TemporalHAN) based on hierarchical attention using a temporal convolutional network (TCN). In particular, we introduce node-level and semantic-level attention into heterogeneous graph neural networks to identify the importance of different levels between nodes. For each snapshot, we first utilise a new random walk algorithm (NRWA) to collect strongly connected heterogeneous neighbours for each node of different types and group them by node types. In addition, the algorithm utilises a damping factor to ensure that the more recent snapshots allocate more random walk steps. We then utilise node-level and semantic-level attention to learn the importance between a node and its random walk neighbour for a specific node type and learn the importance of different node-type for this node, respectively. Finally, we adopt TCN to capture the evolution information between snapshots. Experimental results on relationship prediction and node classification reveal that the TemporalHAN is competitive against diverse state-of-the-art approaches. Our code is available at https://github.com/Legendary-L/THAN.

An approach to the Second Quantum Revolution for secondary school students: the case of the random walk algorithm

The teaching of quantum technologies has now become a leading topic and is at the heart of numerous international programs (e.g., quantum flagship, National Quantum Initiative, UK national quantum technology program) with the aim of widening the workforce and preparing the next generations of experts. In the present contribution, we present an approach for teaching the Second Quantum Revolution to secondary school students that we developed in recent years. The approach aims to emphasize the ongoing revolution as first of all a cultural revolution and its intrinsic interdisciplinary character. As an emblematic case of some aspects of the approach, we present an activity on the classical and quantum random walk algorithms whose main aims are to recontextualize some basic quantum concepts (quantum state, state manipulation/evolution, measurement, and entanglement) in the case of the algorithm and to reflect on the main differences between the classical and quantum case in terms of logic behind its functioning as well as from an epistemological perspective.

Adaptive cuckoo search algorithm based fuzzy C means clustering with random walker algorithm for liver segmentation using CT images

Adaptive cuckoo search algorithm based fuzzy C means clustering with random walker algorithm for liver segmentation using CT images

Weighted Jump in Random Walk graph sampling

Graph sampling is a challenging problem in network analysis due to the complex structures of networks. Currently, a series of graph sampling algorithms based on random walks achieve good results in graph sampling tasks. However, the existing methods often reduce the conductance of graphs, causing the sampler to stay in the same node for a long time. This results in undersampling. In this paper, we propose a novel Weighted Jump Random Walk (WJRW) algorithm to generate representative samples. We design a parameter in the WJRW algorithm that can adjust the proportions of random walk and random jump in every step. According to the issue of repeated sample nodes in the Generalized Maximum Degree (GMD) method and the problem of large deviations in the Simple Random Walk (SRW), WJRW addresses the weaknesses of the GMD method and enhances the diffusion of a random walker on graphs, leading to a more representative sample. Then, WJRW addresses the issue of large deviations in SRW and enhances the efficiency of the unbiased estimator. By generating smoother stationary distributions. Numerical experiments with extensive real-world networks verify that our method achieves higher accuracy than classical and state-of-the-art methods in estimating distribution estimation.

Predicting abiotic stress-responsive miRNA in plants based on multi-source features fusion and graph neural network

BackgroundMore and more studies show that miRNA plays a crucial role in plants' response to different abiotic stresses. However, traditional experimental methods are often expensive and inefficient, so it is important to develop efficient and economical computational methods. Although researchers have developed machine learning-based method, the information of miRNAs and abiotic stresses has not been fully exploited. Therefore, we propose a novel approach based on graph neural networks for predicting potential miRNA-abiotic stress associations.ResultsIn this study, we fully considered the multi-source feature information from miRNAs and abiotic stresses, and calculated and integrated the similarity network of miRNA and abiotic stress from different feature perspectives using multiple similarity measures. Then, the above multi-source similarity network and association information between miRNAs and abiotic stresses are effectively fused through heterogeneous networks. Subsequently, the Restart Random Walk (RWR) algorithm is employed to extract global structural information from heterogeneous networks, providing feature vectors for miRNA and abiotic stress. After that, we utilized the graph autoencoder based on GIN (Graph Isomorphism Networks) to learn and reconstruct a miRNA-abiotic stress association matrix to obtain potential miRNA-abiotic stress associations. The experimental results show that our model is superior to all known methods in predicting potential miRNA-abiotic stress associations, and the AUPR and AUC metrics of our model achieve 98.24% and 97.43%, respectively, under five-fold cross-validation.ConclusionsThe robustness and effectiveness of our proposed model position it as a valuable approach for advancing the field of miRNA-abiotic stress association prediction.

Prediction of MicroRNA-Disease Potential Association Based on Sparse Learning and Multilayer Random Walks.

More and more studies have shown that microRNAs (miRNAs) play an indispensable role in the study of complex diseases in humans. Traditional biological experiments to detect miRNA-disease associations are expensive and time-consuming. Therefore, it is necessary to propose efficient and meaningful computational models to predict miRNA-disease associations. In this study, we aim to propose a miRNA-disease association prediction model based on sparse learning and multilayer random walks (SLMRWMDA). The miRNA-disease association matrix is decomposed and reconstructed by the sparse learning method to obtain richer association information, and at the same time, the initial probability matrix for the random walk with restart algorithm is obtained. The disease similarity network, miRNA similarity network, and miRNA-disease association network are used to construct heterogeneous networks, and the stable probability is obtained based on the topological structure features of diseases and miRNAs through a multilayer random walk algorithm to predict miRNA-disease potential association. The experimental results show that the prediction accuracy of this model is significantly improved compared with the previous related models. We evaluated the model using global leave-one-out cross-validation (global LOOCV) and fivefold cross-validation (5-fold CV). The area under the curve (AUC) value for the LOOCV is 0.9368. The mean AUC value for 5-fold CV is 0.9335 and the variance is 0.0004. In the case study, the results show that SLMRWMDA is effective in inferring the potential association of miRNA-disease.

Color image encryption algorithm based on quantum random walk and multiple reset scrambling* *This work was supported by the Open Fund of Advanced Cryptography and System Security Key Laboratory of Sichuan Province (Grant No. SKLACSS–202208), the Science and Technology Research Program of Chongqing Municipal Education Commission (Grant no. KJZD-M202000501) and National Natural Science Foundation of China (No.61772295).

In order to address the issues of privacy breaches, data tampering, and security threats in the realm of image encryption and information security, it is essential to ensure the confidentiality, integrity, and reliability of image data. In this paper, we propose a novel hybrid color image encryption scheme that increases the complexity of the key space and provides better data integrity protection by combining the quantum random walk and SHA-256 algorithm. At the same time, three chaotic systems were combined, and multiple XOR operations were introduced to effectively scramble the spatial and pixel levels of the color image. Through numerous simulation experiments and security analyses, the results demonstrate that the solution exhibits efficient encryption performance, excellent resistance to attacks, and outstanding privacy protection capabilities.

Link prediction in multilayer networks using weighted reliable local random walk algorithm

The expansion of online social networks has led to attracting a significant number of users. Meanwhile, the analysis of these networks has raised new perspectives and challenges due to the large amount of information. One of these challenges in the analysis of social networks is the problem of link prediction. The purpose of this problem is to find links that have not yet been observed, but may exist in the future. There are many solutions for link prediction on monoplex networks. However, many real social networks model communication in multiple layers. In recent years, link prediction in multilayer social networks has attracted the attention of many researchers. A solution for multilayer networks involves taking into account the information of all layers to make predictions for a target layer. Among the link prediction techniques based on monoplex networks, local random walk is very popular. This paper presents an extended version of this technique for multilayer social networks considering reliable routes and weighted links. Our proposed algorithm as Weighted Reliable Local Random Walk (WRLRW) simultaneously uses interlayer and intralayer information to predict links in a target layer of a multilayer network. WRLRW adjusts link weights based on multimodal and topological features and forms reliable routes between users by multiplying link weights. We demonstrated the effectiveness of WRLRW to address the link prediction problem on real-world multilayer networks through numerical simulations. WRLRW has better performance compared to link prediction techniques based on monoplex networks as well as state-of-the-art methods based on multilayer networks. According to the results of the experiments, WRLRW obtains 3.2% and 2.5% better average f-measure compared to SEM-Path (Structural, Ego-Paths and Meta-Paths features) and MLRW (Multiplex Local Random Walk), respectively.

An anti-greedy random walk algorithm for heat exchanger network synthesis

Heat exchanger network (HEN) synthesis is a vibrant research field in process system engineering, with substantial contributions to energy conservation and emissions reduction initiatives. The optimal design of a heat exchanger network is not an easy task due to the abundance of local optima in the solution space caused by the non-linear, non-convex, and discontinuous nature of the problem. Generally, several heuristic algorithms employ a greedy evolutionary mechanism, optimize through greedily accepting the decrease in the objective function, and converge to obtain the optimal solution. The Random Walk algorithm has a simple evolutionary mechanism, is prone to mutation, and exhibits high flexibility. However, the algorithm's inherent persistent greediness in searching restrict the scope of the search. Thus, this paper proposes an anti-greedy concept based on the Random Walk method to serve as the basis of a new synthesis approach called the Anti-greedy Random Walk algorithm. Two strategies are proposed in the algorithm, which broaden the solution domain by slowing down rapid unit reduction and accepting imperfect solutions, respectively. One strategy is to thoroughly search for the integer and continuous variables of the HEN problem by covering a much larger search space. Another is to escape the local extrema and move forward to discover more possibilities. Quantitative data demonstrates the algorithm's ability to avoid the local extrema and enhance the search effectiveness. Three different scales of classical cases are used in this work and the obtained results are superior to the published ones.

A hybrid approach for automatic segmentation and classification to detect tuberculosis.

Tuberculosis (TB) remains a significant global infectious disease, posing a considerable health threat, particularly in resource-constrained regions. Due to diverse datasets, radiologists face challenges in accurately diagnosing TB using X-ray images. This study aims to propose an innovative approach leveraging image processing techniques to enhance TB diagnostic accuracy within the automatic segmentation and classification (AuSC) framework for healthcare. The AuSC of detection of TB (AuSC-DTB) framework comprises several steps: image preprocessing involving resizing and median filtering, segmentation using the random walker algorithm, and feature extraction utilizing local binary pattern and histogram of gradient descriptors. The extracted features are then classified using the support vector machine classifier to distinguish between healthy and infected chest X-ray images. The effectiveness of the proposed technique was evaluated using four distinct datasets, such as Japanese Society of Radiological Technology (JSRT), Montgomery, National Library of Medicine (NLM), and Shenzhen. Experimental results demonstrate promising outcomes, with accuracy rates of 94%, 95%, 95%, and 93% achieved for JSRT, Montgomery, NLM, and Shenzhen datasets, respectively. Comparative analysis against recent studies indicates superior performance of the proposed hybrid approach. The presented hybrid approach within the AuSC framework showcases improved diagnostic accuracy for TB detection from diverse X-ray image datasets. Furthermore, this methodology holds promise for generalizing other diseases diagnosed through X-ray imaging. It can be adapted with computed tomography scans and magnetic resonance imaging images, extending its applicability in healthcare diagnostics.

Network medicine analysis for dissecting the therapeutic mechanism of consensus TCM formulae in treating hepatocellular carcinoma with different TCM syndromes.

Hepatocellular carcinoma (HCC) is a major cause of cancer-related mortality worldwide. Traditional Chinese Medicine (TCM) is widely utilized as an adjunct therapy, improving patient survival and quality of life. TCM categorizes HCC into five distinct syndromes, each treated with specific herbal formulae. However, the molecular mechanisms underlying these treatments remain unclear. We employed a network medicine approach to explore the therapeutic mechanisms of TCM in HCC. By constructing a protein-protein interaction (PPI) network, we integrated genes associated with TCM syndromes and their corresponding herbal formulae. This allowed for a quantitative analysis of the topological and functional relationships between TCM syndromes, HCC, and the specific formulae used for treatment. Our findings revealed that genes related to the five TCM syndromes were closely associated with HCC-related genes within the PPI network. The gene sets corresponding to the five TCM formulae exhibited significant proximity to HCC and its related syndromes, suggesting the efficacy of TCM syndrome differentiation and treatment. Additionally, through a random walk algorithm applied to a heterogeneous network, we prioritized active herbal ingredients, with results confirmed by literature. The identification of these key compounds underscores the potential of network medicine to unravel the complex pharmacological actions of TCM. This study provides a molecular basis for TCM's therapeutic strategies in HCC and highlights specific herbal ingredients as potential leads for drug development and precision medicine.

Vibration-based sensitivity analysis of a flexible steel footbridge using probabilistic walking models

Many civil engineering structures are concerned with the loads induced by human activity. Human live load modeling of walking and running is very relevant to provide comfort and acceptable structural conditions. The dynamic portion of human loads is often inaccurately modeled or neglected, leading to conservative designs or excessive vibrations. The classical model to consider walking load consists of a dynamic Force Model that considers unique walking parameters along the analysis, related to each pedestrian. Biodynamic models are an alternative to consider the Human-Structure Interaction in the analysis, but also do not vary accordingly to the human activities performed during walking. Human pacing varies continuously during walking, governed by intra- and inter-subject variabilities, giving rise to the possibility of probabilistic modeling of human walking parameters, as already proposed in the literature. In this regard, the article proposes an alternative method by using a Random Walk algorithm to incorporate walking intra-subject variability into the Force or Biodynamic models pacing frequencies. The Random Walk approach shows that walking variability can influence the final service assessment of structures and realistically represent human walking scenarios. It is aimed to discuss the RW parameters’ role and its implications on human walking modeling and vibration performance analyses to improve standardized vibration classification by simulating the walking randomness. A sensitivity analysis is performed to define proper probabilistic distributions and the range of values to be used in the random walk algorithm. A study is also conducted with a vibration performance assessment in idealized footbridges with probabilistic walking force and biodynamic models.

Predicting Tacit Coordination Success Using Electroencephalogram Trajectories: The Impact of Task Difficulty

In this study, we aim to develop a machine learning model to predict the level of coordination between two players in tacit coordination games by analyzing the similarity of their spatial EEG features. We present an analysis, demonstrating the model’s sensitivity, which was assessed through three conventional measures (precision, recall, and f1 score) based on the EEG patterns. These measures are evaluated in relation to the coordination task difficulty, as determined by the coordination index (CI). Tacit coordination games are games in which two individuals are requested to select the same option out of a closed set without the ability to communicate. This study aims to examine the effect of the difficulty of a semantic coordination task on the ability to predict a successful coordination between two players based on the compatibility between their EEG signals. The difficulty of each of the coordination tasks was estimated based on the degree of dispersion of the different answers given by the players reflected by the CI. The classification of the spatial distance between each pair of individual brain patterns, analyzed using the random walk algorithm, was used to predict whether successful coordination occurred or not. The classification performance was obtained for each game individually, i.e., for each different complexity level, via recall and precision indices. The results showed that the classifier performance depended on the CI, that is, on the level of coordination difficulty. These results, along with possibilities for future research, are discussed.

Revealing petroleum-water-CO2 emulsion stability by NMR about oil recovery and carbon storage

CO2 injection can simultaneously enhance oil recovery and carbon storage, but the stability of petroleum-water-CO2 emulsion affcet the production process and carbon storage. Emulsion characteristics is typically assessed via Nuclear Magnetic Resonance (NMR). It is essential to develop a theory to unlock the mechanisms of experimentation for harnessing the full potential of NMR in assessing emulsion stability and comprehending complex multiphase systems. The theory integrates experiments and a “quasi-lattice wave” physical model to understand the microscale mechanisms in NMR measurements of crude oil emulsion properties, and uncover the “transfer-transition” mode of 1H quantum excited states, explaining T2 relaxation. This forms a theoretical basis for numerical simulation of CPMG (Carr-Purcell-Meiboom-Gill) experiments by random walk algorithms and the boundary annihilation hypothesis. A strong link between NMR T2 times and various character properties of Petroleum-Water-CO2 Emulsion samples is reflected, such as environmental scale, viscosity, CO2 adsorption, and molecular thermal motion. Additionally, parameter Θ was proposed to characterize emulsion stability, for penetrating the theory presented in this paper, and to provide a scientifically rigorous representation of crude oil emulsion stability. The experiments with CO2 treatment on crude oil emulsions were performed, demonstrating a highly effective theoretical correlation between interfacial tension in crude oil emulsions and CO2 pressure, with an exceptional fit to experimental data (R2 > 0.99). When the pressure of CO2 treatment exceeds a critical threshold (approximately 17 MPa in this study), there is a substantial reduction in viscosity. Furthermore, a comparative assessment of samples before and after CO2 treatment using the parameter Θ demonstrated its reliability as an indicator of emulsion stability. It revealed a significant reduction of over 50% in the linear fit residual variance of Θ before and after subjecting the sample to 10 MPa carbon dioxide treatment, highlighting improved fluctuation stability. The slope of the Θ linear regression function signifies temporal stability. In a sample with 20% water content, the time evolution of the water phase decreases by 3.04%, while the oil phase experiences a substantial 57.77% reduction. Simultaneously, the midpoint of the range of the Θ linear regression function represents the environmental scale. The water phase environment scale increases by 8.72%, while the oil phase environment scale decreases by 8.74%, highlighting spatial evolution patterns.

Automated diagnostic classification of diabetic retinopathy with microvascular structure of fundus images using deep learning method

Multi-class classification is a major concern in the research field, especially in medical image analysis. This work proposes a novel method for automatically segmenting the blood vessels and classifying Diabetic Retinopathy (DR) using fundus images. This system helps ophthalmologists with the early detection and grading of DR diseases. The main aim here is to detect any pathological changes happening in the retinal vascular structure for the development of DR, through which the patients can avoid undergoing expensive scans. The proposed system involves three different stages, including pre-processing, vessel segmentation, and classification. The input images are processed first to eliminate the noise, followed by green channel extraction and enhancement with Contrast Limited Adaptive Histogram Equalization (CLAHE) and gamma correction. Retinal Vascular Structure (RVS) segmentation is a major concern in this work as it is responsible for detecting the different stages of DR by detecting the presence of microaneurysms, haemorrhages, and exudates. The U-Net is used as a base architecture to develop the segmentation model. The contracting path in the U-Net contains four consecutive downsampling and upsampling layers with skip connections. After performing downsampling four times, information on the tiny blood vessels may be missed. Therefore, ResEAD2Net is introduced in this work, where the number of downsampling and upsampling layers is reduced to two instead of four, and two such contracting paths and expansion paths are added to the network. Thus, detailed semantic information can be retained with this structure. Residual blocks are included instead of convolution blocks to increase the computational speed. To include structural connectivity, the segmented output is passed through the proposed Regularized Random Walker (RRW) algorithm, which focuses on the broken blood vessels. Finally, the features are extracted from the vessel structure and passed through the Machine Learning (ML) classifier to predict the DR grading. The proposed method achieves better performance in segmentation with accuracy, sensitivity, specificity, and area under the curve values of 98.07%, 90.24%, 99.01%, and 97.51%, respectively, for the STARE dataset and 97.55, 90.07, 98.01, and 97.64, respectively, for the DRIVE dataset. It is also found that the features obtained from the segmentation and ML classifier outperforms the existing methods in multiclass classification by achieving accuracy, sensitivity, and specificity of 98.88%, 98.91%, and 98.29% with the Messidor-2 dataset.