Phenetic Research Articles

Development of Teaching Performance Evaluation Application for Lecturers Using K-Nearest Neighbor Method with Manhattan Distance Approach

Based on the initial observation, there are several underlying issues that form the basis of this research. The teaching performance evaluation at Padang State University has limitations regarding the courses to be evaluated. Each student can only evaluate 5 (five) courses per semester, where these five courses are randomly selected by the system, allowing each student to evaluate different courses even in the same class. The evaluation of teaching performance at Padang State University is not specific to individual lecturers but to the courses. One course can be taught by several lecturers, so students evaluating the learning cannot provide assessments for each lecturer. This results in each lecturer not having their own performance results. Furthermore, the teaching performance evaluation at Padang State University does not have a classification for the filled evaluations, thus requiring a long time to calculate the final results. This study uses the Research and Development (R&D) method with the 4-D development model consisting of four stages: definition, design, development, and dissemination. The type of data used is primary data obtained from 3 media validators, the evaluation administration of Padang State University, and 46 students. The data analysis technique used is descriptive data analysis to describe the validity and practicality of the developed lecturer performance evaluation application. The results of this development study produced a lecturer performance evaluation with 202 training data, 47 test data, resulting in an accurate system with a precision value of 88.76%, a recall value of 89.93%, and a program accuracy value of 94.04%. The validity results of the web-based learning evaluation conducted by media experts obtained a score of 0.864 with a valid category. The practicality value of using web-based lecturer performance evaluation by students obtained a score of 87.42 with a practical category. Meanwhile, the practicality value obtained by the evaluation administration of Padang State University is 89.33 with a practical category.

Early-Stage ISC Fault Detection for Ship Lithium Batteries Based on Voltage Variance Analysis

With the progressive development of new energy technologies, high-power lithium batteries have been widely used in ship power systems due to their high-power density and low environmental pollution, and they have gradually become one of their main propulsion energy sources. However, the large-scale deployment of lithium batteries has also brought a series of safety problems to ship operations, especially the battery internal short circuit (ISC). Battery ISC faults are very hidden and unpredictable at the initial stage and often fail to be detected in time, ultimately leading to overheating, fire or even an explosion of the ship’s power system. Based on this, this paper proposes a fast and accurate method for early-stage ISC fault location and detection of lithium batteries. Initially, voltage variations across the lithium battery packs are quantified using curvilinear Manhattan distances to pinpoint faulty battery units. Subsequently, the localized characteristics of voltage variance among adjacent batteries are leveraged to detect an early-stage ISC fault. Simulation results indicate that the proposed method can quickly and accurately locate the position of 5 Ω, 10 Ω and 15 Ω ISC faulty batteries within the battery pack, as well as detect the abnormal batteries in a timely manner with considerable sensitivity and reliability.

Velocity-based rack storage location assignment for the unidirectional robotic mobile fulfillment system

Nowadays, the robotic mobile fulfillment system (RMFS) has been increasingly used by online retailers. Compared with traditional picker-to-parts warehouses, the racks of RMFS do not have to return to the same location after picking, thus we can dynamically change their locations, which brings great potential to efficiently fulfill orders. Aiming at minimizing the sum of rack travel distances, the key question is how to reassign a rack to an unoccupied storage location after picking items from the rack. However, the issue involves two challenges for unidirectional RMFS, one is that huge differences may exist between the classical Manhattan distance estimation and the actual distance for the unidirectional aisles in RMFS, and the other one is that we need to account for the frequency of rack moving for multi-item orders. We thus first propose closed-form formulas to optimally estimate the cycle travel distance for each rack. Then, by overcoming the repeated counting issue for multi-item orders, we propose a novel SKU (Stock Keeping Units)-correlation-based algorithm to choose high-velocity racks, which can better fulfill multi-item orders. Finally, embedding the cycle travel distance and SKU-correlation-based velocity, we propose a Velocity-based Rack Storage Location Assignment method (VRSLA) to solve the rack storage location assignment problem by assigning high-velocity racks to the nearest storage locations. Collaborating with a large online retailer in China, we demonstrate the performance of VRSLA by using both small-scale and large-scale datasets. The computational results show that VRSLA not only can achieve near-best solutions compared with an integer programming model solved by Gurobi, but also outperforms four state-of-the-art assignment methods in literature (random, velocity-based class, shortest path, and sale-based) by reducing the rack travel distance up to 43.32%. We also found that the stronger the correlation between SKUs on the racks or the larger the size of the RMFS, the shorter the rack travel distance by the proposed VRSLA method.

A Novel Method for Remaining Useful Life Prediction of RF Circuits Based on the Gated Recurrent Unit–Convolutional Neural Network Model

The remaining useful life (RUL) prediction of RF circuits is an important tool for circuit reliability. Data-driven-based approaches do not require knowledge of the failure mechanism and reduce the dependence on knowledge of complex circuits, and thus can effectively realize RUL prediction. This manuscript proposes a novel RUL prediction method based on a gated recurrent unit–convolutional neural network (GRU-CNN). Firstly, the data are normalized to improve the efficiency of the algorithm; secondly, the degradation of the circuit is evaluated using the hybrid health score based on the Euclidean and Manhattan distances; then, the life cycle of the RF circuits is segmented based on the hybrid health scores; and finally, an RUL prediction is carried out for the circuits at each stage using the GRU-CNN model. The results show that the RMSE of the GRU-CNN model in the normal operation stage is only 3/5 of that of the GRU and CNN models, while the prediction uncertainty is minimized.

Dialectometry-based classification of the Central–Southern Italian dialects

Abstract This paper provides a new classification of Central–Southern Italian dialects using dialectometric methods. All varieties considered are analyzed and cast in a data set where homogeneous areas are evaluated according to a selected list of phonetic features. Using numerical evaluation of these features and the Manhattan distance, a linguistic distance rule is defined. On this basis, the classification problem is formulated as a clustering problem, and a k-means algorithm is used. Additionally, an ad-hoc rule is set to identify transitional areas, and silhouette analysis is used to select the most appropriate number of clusters. While meaningful results are obtained for each number of clusters, a nine-group classification emerges as the most appropriate. As the results suggest, this classification is less subjective, more precise, and more comprehensive than traditional ones based on selected isoglosses.

A unique approach for protein secondary structure comparison under TOPS representation

To unravel the intricate connection between protein function and protein structure, it is imperative to comprehensively evaluate protein secondary structure similarity from various perspectives. While numerous techniques have been suggested for comparing protein secondary structure elements (SSE), there continues to be a substantial need for finding alternative ways of comparing the same. In this paper, Topology of Protein Structure (TOPS) representations of protein secondary structures are considered to offer a new alignment-free method for evaluating similarities/dissimilarities of protein secondary structures. Initially, a two-dimensional numerical representation of the SSE is created, associating each point with a mass reflecting its frequency of occurrence. Then the means of coordinate values are determined by averaging weighted sums, and these mean values are subsequently used to calculate moments-of-inertia. Next, a four-component descriptor is generated out of the eigenvalues of the matrix and the mean values of the represented coordinates. Thereafter, Manhattan distance measure is used to obtain the distance matrix. This is finally applied to obtain the phylogenetic trees under the use of NJ method. SSE considered in the proposed method comprises 36-elements from the Chew-Kedem database giving five different taxa: globin, alpha-beta, tim-barrel, beta, and alpha. Phylogenetic trees were created for these SSE through the application of various methods: Clustal-Omega, LZ-Complexity, SED, TOPS + and TOC, to facilitate comparative analysis. Phylogenetic tree of the proposed method outperformed results of the previous methods when applied to the same SSE. Therefore, the method effectively constructs phylogenetic tree for analyzing protein secondary structure comparison. Communicated by Ramaswamy H. Sarma

Elliptic Inversions in Taxicab Geometry

The goal of this research is to introduce inversion with respect to an ellipse which is a generalization of the classical circular inversion in taxicab plane and to investigate general properties and basic concepts of this transformation in taxicab geometry. The cross ratio is preserved under the elliptic inversion in taxicab plane though this transformation is not an isometry. Thus some properties such as cross ratio and harmonic conjugates of the elliptic inversions in R_T^2 are also studied.

An airborne radar forward-looking imaging algorithm based on multi-scale U-shaped networks

ABSTRACT Radar forward-looking imaging has gained increasingly significance in diverse applications, including battlefield reconnaissance, target surveillance, and precision guidance. While Synthetic Aperture Radar techniques are widely employed to achieve high azimuth resolution, the domain of forward-looking imaging has its challenges. These challenges arise due to the minimal Doppler frequency changes in forward-looking areas and the symmetrical Doppler history about flight path. These constraints negatively impact the radar’s ability to accurately detect and image targets in forward-looking areas. Consequently, there has been a growing interest in the development of novel approaches to these problems. Due to the application of deep learning, image super-resolution methods based on convolutional neural networks (CNNs) have received extensive attention and demonstrated good performance. This avenue of research provides fresh insights into improving the azimuth resolution of radar imaging. In this letter, we propose a novel convolutional neural network, called MSUS-Net (Multi-scale U-shaped Network), for radar forward-looking imaging. It combines asymmetric feature fusion with a unique connection structure to replace traditional layers. Moreover, different from conventional image restoration networks that use content loss, we use the L1 distance in the frequency domain, restoring missing high-frequency details. Simulated and real radar data validate that this approach offers robust recovery and better imaging performance.

Multi wavelength optical image feature fusion method for improving the reconstruction effect of cerebral hematoma

Diffuse optical tomography (DOT) under medical image guidance can be used to reconstruct images of subdural hematomas for monitoring purposes. However, when the shape of hematoma is irregular or the signal to noise ratio(SNR) of input signal is low, the accuracy and robustness of this method both decrease. In order to alleviate this situation and improve the reconstruction effect of cerebral hematoma images, a multi wavelength optical image feature fusion method was proposed in this paper. Experimental results show that the mean relative volume error (VRE) of the hematoma reconstruction model based on the multi-wavelength optical image feature fusion method is only 0.66 %, and the mean value of the average Manhattan distance (AMD) between the reconstructed absorption coefficient and the true absorption coefficient is 0.0052 mm−1. Compared with the model using single-wavelength optical image mapping feature extraction method, the multi-wavelength optical image feature fusion method reduces the average VRE of the model by 16.5 %, and the average AMD by 16.1 %. And compared with the model directly inputting single-wavelength optical information, its average VRE and average AMD decreased by 80.0 % and 27.8 %. This method has clinical application significance for the monitoring of subdural hematomas.

A Hierarchical Block Distance Model for Ultra Low-Dimensional Graph Representations

A Hierarchical Block Distance Model for Ultra Low-Dimensional Graph Representations

Comparison of K-NN, SVM, and Random Forest Algorithm for Detecting Hoax on Indonesian Election 2024

During the year 2022, The Indonesian National Police (POLRI) received 113 reports related to the spread of hoax news related to 2024 Indonesian Election (PEMILU). There are still relatively few hoax detection tools that already exist in Indonesia. This research creates a system that can detect hoax news in Indonesian tweets about the Indonesian Election (PEMILU) 2024 by comparing three methods, namely K-NN, SVM, and Random Forest. The process of labeling (create model) using validation on ground truth data, namely cekfakta.tempo, cekfakta.kompas, and turnbackhoax.id. In this research, we also check the differences between different types of distance measurements in applying the K-NN algorithm. The method used for feature extraction in this research is TF-IDF. The results of experiments show that the highest accuracy results are obtained using the SVM and K-NN algorithms with distance measurements using Euclidean Distance, which is 86.36%. The best precision value is obtained using the K-NN algorithm with distance measurements using Manhattan Distance, which is 86.95%.

Offline Handwriting Writer Identification using Depth-wise Separable Convolution with Siamese Network

Offline handwriting writer identification has significant implications for forensic investigations and biometric authentication. Handwriting, as a distinctive biometric trait, provides insights into individual identity. Despite advancements in handcrafted algorithms and deep learning techniques, the persistent challenges related to intra-variability and inter-writer similarity continue to drive research efforts. In this study, we build on well-separated convolution architectures like the Xception architecture, which has proven to be robust in our previous research comparing various deep learning architectures such as MobileNet, EfficientNet, ResNet50, and VGG16, where Xception demonstrated minimal training-validation disparities for writer identification. Expanding on this, we use a model based on similarity or dissimilarity approaches to identify offline writers' handwriting, known as the Siamese Network, that incorporates the Xception architecture. Similarity or dissimilarity measurements are based on the Manhattan or L1 distance between representation vectors of each input pair. We train publicly available IAM and CVL datasets; our approach achieves accuracy rates of 99.81% for IAM and 99.88% for CVL. The model was evaluated using evaluation metrics, which revealed only two error predictions in the IAM dataset, resulting in 99.75% accuracy, and five error predictions for CVL, resulting in 99.57% accuracy. These findings modestly surpass existing achievements, highlighting the potential inherent in our methodology to enhance writer identification accuracy. This study underscores the effectiveness of integrating the Siamese Network with depth-wise separable convolution, emphasizing the practical implications for supporting writer identification in real-world applications.

Penerapan Algoritma K-Means untuk Klasterisasi Produksi Budidaya Perikanan Provinsi Sulawesi Utara

Capture fisheries production is decreasing due to natural resources or weather conditions, so other resources are needed to support fisheries production. One alternative is to increase the production of aquaculture commodities through seawater, freshwater, or brackish water cultivation. Various potentials for developing aquaculture have been developed in various regions, including the North Sulawesi region. Grouping aquaculture commodity production according to container type is very important to maintain and increase aquaculture production. This research aims to cluster aquaculture production in the North Sulawesi area using the K-Means method using Euclidean, Manhattan, and Minkowsky distances. The results of the research obtained three clusters, namely the first cluster, the region with the highest production of pond container types, namely the Sitaro Islands, and for second cluster consisting of 13 regions that have variations in production ranging from low to high for the types of floating net containers and ponds, while third cluster is the region Bitung with moderate production for pond types. It is hoped that this research can help related agencies to create policies to increase the production potential of aquaculture.

Optimal Temperature Calculation for Multicriteria Optimization of the Hydrogenation of Polycyclic Aromatic Hydrocarbons by NSGA-II Method

Introduction. Multicriteria optimization, taking into account contradicting criteria, is used to improve production efficiency, reduce costs, improve product quality and environmental safety of processes. The literature describes the application of multicriteria optimization for production purposes, including the selection of reaction conditions and improvement of technological processes. In the presented paper, the object of research is the process of hydrogenation of polycyclic aromatic hydrocarbons (PAH) in the production of high-density fuels. To determine the optimal conditions of the process, the problem of multicriteria optimization based on the kinetic model is solved. The criteria include maximizing the yield of targeted naphthenes and conversion of feedstock. The research objective is to create a program implementing the multicriteria optimization non-dominated sorting genetic algorithm-II (NSGA-II). Due to this, it is possible to calculate the optimal temperature for the PAH hydrogenation process on the basis of the kinetic model.Materials and Methods. The NSGA-II genetic algorithm was used to solve the multicriteria optimization problem. Modified parental and survival selection within the Pareto front was also used. If it was necessary to divide the front, solutions based on the Manhattan distance between them were selected. The program was implemented in Python.Results. In the system of ordinary nonlinear differential equations of chemical kinetics, the concentration was designated yi, the conditional contact time of the reaction mixture with the catalyst — τ. The system was solved for the hydrogenation reaction of polycyclic aromatic hydrocarbons. The calculations showed that at τ = 0 y1(0) = 0.025; y2(0) = 0.9; y6(0) = 0.067; y9(0) = 0.008; yi(0) = 0, i = 3–5, 7, 8, 10–20; Q(0) = 1. The process temperature was considered as a control parameter according to two optimality criteria: maximizing the yield of target naphthenes (f1) at the end of the reaction, and maximizing the conversion of feedstock (f2). Values f1 were in the range of 0.43–0.79; conversion — 0.01–0.03; temperature — 200–300 K. The growth of temperature was accompanied by an increase in the yield of target naphthenes and a decrease in the conversion of feedstock. Each solution obtained was not an unimprovable one. When modeling the process of hydrogenation of PAH, an algorithm was launched with a population size of 100 and a number of generations of 100. A program implementing the NSGA-II algorithm was developed. The optimal set of values of the PAH hydrogenation reaction temperature was calculated, which made it possible to obtain unimprovable values of the optimality criteria — maximizing the yield of target naphthenes and conversion of feedstock.Discussion and Conclusion. The NSGA-II algorithm is effective for solving the problem of non-dominance, and deriving the optimal solution for all criteria. Future research should be devoted to the selection of optimal algorithm parameters to increase the speed of the solution. Based on the obtained theoretical optimal conditions of the PAH hydrogenation reaction, it is possible to implement the process in industry

Down-sampling in diffusion MRI: a bundle-specific DTI and NODDI study.

Multi-shell diffusion Magnetic Resonance Imaging (dMRI) data has been widely used to characterise white matter microstructure in several neurodegenerative diseases. The lack of standardised dMRI protocols often implies the acquisition of redundant measurements, resulting in prolonged acquisition times. In this study, we investigate the impact of the number of gradient directions on Diffusion Tensor Imaging (DTI) and on Neurite Orientation Dispersion and Density Imaging (NODDI) metrics. Data from 124 healthy controls collected in three different longitudinal studies were included. Using an in-house algorithm, we reduced the number of gradient directions in each data shell. We estimated DTI and NODDI measures on six white matter bundles clinically relevant for neurodegenerative diseases. Fractional Anisotropy (FA) measures on bundles where data were sampled at the 30% rate, showed a median L1 distance of up to 3.92% and a 95% CI of (1.74, 8.97)% when compared to those obtained at reference sampling. Mean Diffusivity (MD) reached up to 4.31% and a 95% CI of (1.60, 16.98)% on the same premises. At a sampling rate of 50%, we obtained a median of 3.90% and a 95% CI of (1.99, 16.65)% in FA, and 5.49% with a 95% CI of (2.14, 21.68)% in MD. The Intra-Cellular volume fraction (ICvf) median L1 distance was up to 2.83% with a 95% CI of (1.98, 4.82)% at a 30% sampling rate and 3.95% with a 95% CI of (2.39, 7.81)% at a 50% sampling rate. The volume difference of the reconstructed white matter at reference and 50% sampling reached a maximum of (2.09 ± 0.81)%. In conclusion, DTI and NODDI measures reported at reference sampling were comparable to those obtained when the number of dMRI volumes was reduced by up to 30%. Close to reference DTI and NODDI metrics were estimated with a significant reduction in acquisition time using three shells, respectively with: 4 directions at a b value of 700 s/mm2, 14 at 1000 s/mm2, and 32 at 2000 s/mm2. The study revealed aspects that can be important for large-scale clinical studies on bundle-specific diffusion MRI.

SHORTEST PATH DISTANCE AND HAUSDORFF DIMENSION OF SIERPINSKI NETWORKS

In this paper, we will study the geometric structure on the Sierpinski networks which are skeleton networks of a connected self-similar Sierpinski carpet. Under some suitable condition, we figure out that the renormalized shortest path distance is comparable to the planar Manhattan distance, and obtain the Hausdorff dimension of Sierpinski networks.

Multi-dimensional multi-round minimum cost consensus models with iterative mechanisms involving reward and punishment measures

In order to address the issues of unfairness in consensus and subjectivity bias in data, while further expanding the traditional minimum cost consensus model (MCCM) in terms of opinion dimension and consensus round, we propose the multi-dimensional MCCM (MD-MCCM) and multi-round MCCM (MR-MCCM). Subsequently, we introduce the concepts and definitions pertaining to these two variants of consensus and formulate their respective consensus model frameworks. To benefit from the advantages of both, we integrate these two types of consensus models, thereby introducing the concept of multi-dimensional multi-round minimum cost consensus. To address the issues of unfairness and subjectivity in the consensus-reaching process, we propose two specific reward and punishment measures and devise the multi-dimensional multi-round consensus iterative mechanisms. These mechanisms facilitate an iterative update of decision-makers’ opinions, weights, and unit costs within the feedback process. Based on this foundation, we construct the multi-dimensional multi-round MCCM (MDMR-MCCM) involving Euclidean and Manhattan distance measures, and their corresponding algorithm and consensus reaching process are presented. Finally, a group decision problem concerning site selection validates the rationality and effectiveness of our models. Moreover, a comparative analysis of results obtained using different models is provided.

Foreseeing Reconstruction Quality of Gradient Inversion: An Optimization Perspective

Gradient inversion attacks can leak data privacy when clients share weight updates with the server in federated learning (FL). Existing studies mainly use L2 or cosine distance as the loss function for gradient matching in the attack. Our empirical investigation shows that the vulnerability ranking varies with the loss function used. Gradient norm, which is commonly used as a vulnerability proxy for gradient inversion attack, cannot explain this as it remains constant regardless of the loss function for gradient matching. In this paper, we propose a loss-aware vulnerability proxy (LAVP) for the first time. LAVP refers to either the maximum or minimum eigenvalue of the Hessian with respect to gradient matching loss at ground truth. This suggestion is based on our theoretical findings regarding the local optimization of the gradient inversion in proximity to the ground truth, which corresponds to the worst case attack scenario. We demonstrate the effectiveness of LAVP on various architectures and datasets, showing its consistent superiority over the gradient norm in capturing sample vulnerabilities. The performance of each proxy is measured in terms of Spearman's rank correlation with respect to several similarity scores. This work will contribute to enhancing FL security against any potential loss functions beyond L2 or cosine distance in the future.

Progressively Knowledge Distillation via Re-parameterizing Diffusion Reverse Process

Knowledge distillation aims at transferring knowledge from the teacher model to the student one by aligning their distributions. Feature-level distillation often uses L2 distance or its variants as the loss function, based on the assumption that outputs follow normal distributions. This poses a significant challenge when distribution gaps are substantial since this loss function ignores the variance term. To address the problem, we propose to decompose the transfer objective into small parts and optimize it progressively. This process is inspired by diffusion models from which the noise distribution is mapped to the target distribution step by step. However, directly employing diffusion models is impractical in the distillation scenario due to its heavy reverse process. To overcome this challenge, we adopt the structural re-parameterization technique to generate multiple student features to approximate the teacher features sequentially. The multiple student features are combined linearly in inference time without extra cost. We present extensive experiments performed on various transfer scenarios, such as CNN-to-CNN and Transformer-to-CNN, that validate the effectiveness of our approach.

Transformer-based multimodal change detection with multitask consistency constraints

Change detection plays a fundamental role in Earth observation for analyzing temporal iterations over time. However, recent studies have largely neglected the utilization of multimodal data that presents significant practical and technical advantages compared to single-modal approaches. This research focuses on leveraging pre-event digital surface model (DSM) data and post-event digital aerial images captured at different times for detecting change beyond 2D. We observe that the current change detection methods struggle with the multitask conflicts between semantic and height change detection tasks. To address this challenge, we propose an efficient Transformer-based network that learns shared representation between cross-dimensional inputs through cross-attention. It adopts a consistency constraint to establish the multimodal relationship. Initially, pseudo-changes are derived by employing height change thresholding. Subsequently, the L2 distance between semantic and pseudo-changes within their overlapping regions is minimized. This explicitly endows the height change detection (regression task) and semantic change detection (classification task) with representation consistency. A DSM-to-image multimodal dataset encompassing three cities in the Netherlands was constructed. It lays a new foundation for beyond-2D change detection from cross-dimensional inputs. Compared to five state-of-the-art change detection methods, our model demonstrates consistent multitask superiority in terms of semantic and height change detection. Furthermore, the consistency strategy can be seamlessly adapted to the other methods, yielding promising improvements.