Nearest-neighbor Technique Research Articles

A Novel Energy-Aware Path Planning by Autonomous Underwater Vehicle in Underwater Wireless Sensor Networks

Wireless sensor networks can monitor the environment to detect anomalies and reduce the risk of maritime traffic. Energy is necessary for low-power conditions where wireless sensor networks are used. Ensuring the lifespan of energy constraints and providing continuous environmental observation, data collecting, and communication requires management. Battery replacement and energy consumption issues can be resolved with path planning and energy-efficient autonomous underwater vehicle charging for sensor nodes. The nearest neighbour technique is used in this study to solve the energy-aware path planning problem of an autonomous underwater vehicle. Path planning simulations show that the nearest neighbour strategy converges faster and produces a better result than the genetic algorithm. We develop robust and energy-efficient path-planning algorithms that efficiently acquire sensor data while consuming less energy, allowing the monitoring system to respond to anomalies more rapidly. Increased sensor connectivity lowers energy usage and increases network longevity. This study also considers the situation when it is recommended to avoid taking direct travel paths between particular node pairs for a variety of reasons. This recommendation is considered in this study. We present a strategy based on a modified Nearest Neighbour-based Approach from the Nearest Neighbour method to address this more challenging scenario. The direct pathways between such nodes are constrained within the context of this technique. The modified version of Nearest Neighbor-based approach performs well even in that particular situation.

TPANet: A novel triple parallel attention network approach for remaining useful life prediction of lithium-ion batteries

Accurate remaining useful life (RUL) prediction for lithium-ion batteries (LIBs) is important for proper equipment operation. Among the numerous existing studies on battery research, data-driven approaches have gained significant attention because they obviate the need for complex chemical and physical modeling of battery processes. However, in most current data-driven methods, the sliding window size values are determined empirically, which affects both the time required for model prediction and the prediction accuracy. To address this issue, this paper proposes a two-stage prediction method for battery capacity aging trajectories. In the first stage, a false nearest neighbor (FNN) technique is utilized to infer the sliding window size of the battery, reducing the error associated with manually selecting the sliding window size. In the second stage, a novel triple parallel attention network (TPANet) based on convolutional neural networks (CNNs) and attention units is developed, which extracts the input battery capacity degradation features through a parallel mechanism. The introduction of attention units in each parallel branch allows the model to enhance the capture of capacity regeneration phenomena as well as long-term dependence on input data. The proposed approach is validated on two publicly available datasets as well as a battery developed by ourselves. Diverse simulation results demonstrate the ability of the proposed approach to accurately predict the RUL of LIBs in a short time with broad generalization capabilities.

Stone decision engine accurately predicts stone removal and treatment complications for shock wave lithotripsy and laser ureterorenoscopy patients.

Kidney stones form when mineral salts crystallize in the urinary tract. While most stones exit the body in the urine stream, some can block the ureteropelvic junction or ureters, leading to severe lower back pain, blood in the urine, vomiting, and painful urination. Imaging technologies, such as X-rays or ureterorenoscopy (URS), are typically used to detect kidney stones. Subsequently, these stones are fragmented into smaller pieces using shock wave lithotripsy (SWL) or laser URS. Both treatments yield subtly different patient outcomes. To predict successful stone removal and complication outcomes, Artificial Neural Network models were trained on 15,126 SWL and 2,116 URS patient records. These records include patient metrics like Body Mass Index and age, as well as treatment outcomes obtained using various medical instruments and healthcare professionals. Due to the low number of outcome failures in the data (e.g., treatment complications), Nearest Neighbor and Synthetic Minority Oversampling Technique (SMOTE) models were implemented to improve prediction accuracies. To reduce noise in the predictions, ensemble modeling was employed. The average prediction accuracies based on Confusion Matrices for SWL stone removal and treatment complications were 84.8% and 95.0%, respectively, while those for URS were 89.0% and 92.2%, respectively. The average prediction accuracies for SWL based on Area-Under-the-Curve were 74.7% and 62.9%, respectively, while those for URS were 77.2% and 78.9%, respectively. Taken together, the approach yielded moderate to high accurate predictions, regardless of treatment or outcome. These models were incorporated into a Stone Decision Engine web application (http://peteranoble.com/webapps.html) that suggests the best interventions to healthcare providers based on individual patient metrics.

Mapping Junior High School Education Facilities in Bogor: A Comprehensive Study

Equitable distribution of educational facilities is very important because every citizen has the right to a decent and quality education as mandated in Law No. 24/2007 on minimum standards of educational facilities and infrastructure. Junior high school education is a bridge between elementary school and senior high school to university. This study and mapping of junior high school education facilities in Bogor Regency is intended to identify the distribution pattern and needs of school facilities and infrastructure at the junior high school level in Bogor Regency. The objectives of this study are 1) To identify the distribution pattern of junior high schools in Bogor Regency; 2) Identify the availability of junior high school facilities and infrastructure in Bogor Regency; and 3) Identify the availability of educational service facilities at junior high school level. Based on the data, analysis was conducted using qualitative and quantitative analysis methods. Data processing using the nearest neighbor technique (average nearest neighbor analysis) is to determine the distribution pattern of junior high school education facilities. Meanwhile, to determine the range of services using Geographic Information System (GIS) buffering analysis techniques. so that the results obtained, 1) The distribution pattern of junior high school facilities in Bogor Regency is clustered with a Critical Value (Z-Score) value of -21.829520, a Nearest Neighbor Ratio value of 0.658887 and an average distance between junior high school education facilities of 963.7 meters; 2) The range of services of junior high school facilities in Bogor Regency is dominated by road lanes.Keywords: study, mapping, distribution pattern, education facilities, junior high school.

Recurrent Convolution Neural Network for Space Debris and Asteroids Detection and Classification on the Monitored Satellite Information Using Deep Space Network

Satellites are used for many monitoring, prediction and forecasting application operation in space , hence it has to be continuously monitored and controlled against the catastrophic damage due to collision of space debris and asteroids debris in the space. Timely and efficient Controlling of satellite will eliminate the heavy irreparable damages as number of space debris and asteroids debris is continuously increasing in the space. Hence monitoring of the satellites increases the sustainability and its safety navigation in the space operations. In order to achieve this objective, deep space network (Satellite ground station ) has to be designed as space situational awareness mechanism on basis of the detection and classification of the space objects towards providing optimal path for satellite trajectory with high processing speed and reduction of memory in data processing. In this work, a new deep learning architecture entitled as optimized recurrent convolution neural network is designed to satellite ground station towards detection and classification of space debris and asteroids debris. Initially architecture is interfaced with feature extraction technique named as principle component analysis to extract the space features and those extracted feature were detected and classified using deep learning model. In this work, convolution neural network composed of convolution layer is used to filter the discrete features, max pooling layer for downsizing the features and represent the feature map and finally fully connected layer to detect and classify the space debris and predict the optimal path for satellite trajectory on basis of the velocity of the debris in the particular orbit. Experimental results of the proposed model has been evaluated using accuracy and precision metric. Further design is implemented and verified using Xilinx Software. The performance analysis of the proposed model yields better accuracy on the classifying the monitored information on comparing with conventional approaches such as support vector machine and K- Nearest Neighbour Technique. Keywords: Satellite Collision Analysis, Deep learning, Space Debris , Asteroid Debris, Convolution Neural Network, Principle Component analysis , Deep Space Network

Subjective Cognitive Decline Prediction on Imbalanced Data Using Data-Resampling and Cost-Sensitive Training Methods

Dementia is a growing public health and health-care problem. Given there is no cure for dementia, its early detection is critical. This would help in introducing rehabilitation and interventions earlier in the disease trajectory, potentially slowing down the cognitive decline. Subjective Cognitive Decline (SCD) indicative of memory loss/confusion is considered among the earliest symptomatic indicators of dementia, preceding Mild Cognitive Impairment (MCI). Therefore, its prediction is relevant for early detection of dementia. Moreover, SCD prediction is under-explored and not many insights currently exist. In this work, we model SCD risk on real-world sparse and imbalanced data using Ensemble models. Two techniques for handling imbalanced learning are compared: Data Resampling Methods and Cost-Sensitive Training. AdaBoost classifier based on hybrid data obtained from Edited Nearest Neighbor (under-sampling) and Synthetic Minority Oversampling Technique - Numerical and Categorical (over-sampling) is su- perior in predicting SCD. For the final model based on this combination, the Sensitivity is increased from 3% in the imbalanced data to 58%, without much loss in Specificity (70%). The work is novel because it focuses on pre-MCI prediction while presenting and comparing techniques to handle the inherent imbalanced distribution in the data.

Differentiating Parkinson's Disease from other Neuro Diseases and Diagnosis using Deep Learning with Nature Inspired Algorithms and Ensemble Learning

A highly accurate automated approach using a deep structured neural network was proposed to detect Parkinson's disease through voice samples. This cost-effective and non-invasive method aims to enhance diagnostic precision and assess the disease's stage of progression. The study addresses two classification problems: binary and multi-category classification. The binary classification deep structured neural network achieved a diagnostic accuracy of 97.6%. For multiclass classification, a Deep Convolution Neural Network (DCNN) and a K- nearest neighbor technique for benchmarking, were utilized using a shared database. Nature Inspired Algorithm (NIA)-Genetic Algorithm was employed to optimize the initial feature set. The suggested deep structure neural network demonstrated a 93.7% accuracy in estimating the disease's stage, showing promising results. Further investigations are encouraged to explore the model's adaptability and pursue improved outcomes.

Security analysis of the Classic McEliece, HQC and BIKE schemes in low memory

With the advancement of NIST PQC standardization, three of the four candidates in Round 4 are code-based schemes, namely Classic McEliece, HQC and BIKE. Currently, one of the most important tasks is to further analyze their security levels for the suggested parameter sets. At PKC 2022, Esser and Bellini restated the major information set decoding (ISD) algorithms by using nearest neighbor technique and then applied these ISD algorithms to estimate the bit security of Classic McEliece, HQC and BIKE under the suggested parameter sets. However, all major ISD algorithms consume a large amount of memory, which in turn affects their time complexities. In this paper, we theoretically reestimate the security levels of the parameter sets suggested by these three schemes in low memory by applying K-list sum algorithms to ISD framework. Compared with Esser–Bellini’s algorithms, our results achieve the best gains for Classic McEliece, HQC, and BIKE, with reductions in security levels of 11.09, 12.64, and 12.19 bits, respectively.

A Backbone-Network-Construction-Based Multi-AUV Collaboration Source Location Privacy Protection Algorithm in UASNs

Underwater acoustic sensor networks (UASNs) are effective instruments for monitoring marine environments and surveying seabed resources, it is important to improve their security protection, including source location privacy protection. Numerous strategies have been presented by researchers to strengthen location privacy, however, the majority of these plans have expensive energy costs. Therefore, a backbone network construction-based multi-autonomous underwater vehicle (AUV) collaboration source location privacy protection (BNCSLP) algorithm has been enhanced to address this issue. First, the nodes in the network are split into various clusters, and an entire network is segmented into various regions. The backbone network is built using clusters that house the source. To prevent the adversary’s tracking, the AUV alternately chooses alternative backbone nodes and relays the source and fake data. Then, the cluster head determines whether to update the clusters by calculating how similar the data are with nearby clusters. The nearest neighbor technique is used by the updated cluster to anticipate the data and to reduce the energy utilization of forwarding the data packets, resulting in that there is less chance of data packets being intercepted and the source location being revealed. Finally, the AUV re-plans the trajectory, which shortens the AUV’s travelling path because only fewer cluster heads need to be accessed, decreasing the time it takes for data to be transmitted.

Analyzing spatial distributions and alignments of pitted cone features in Utopia Planitia on Mars

Martian geomorphology and surface features provide links to understanding past geologic processes such as fluid movement, local and regional tectonics, and feature formation mechanisms. Pitted cones are common features in the northern plains basins of Mars. They have been proposed to have formed from upwelling volatile-rich fluids, such as magma or water-sediment slurries. In this study, we map the spatial distributions of pitted cone features across the Utopia Planitia (UP) basin. Using the Average Nearest Neighbor technique, we find that pitted cone features appear to be clustered across the basin, occurring in a narrow band around the circumferential rim of UP. Additionally, we find that pitted cone features also appear to be aligned in chains that are sub-parallel with the basin margin. Parallel bands of cones generally follow elevation contours of the UP basin, which suggests elevation or a correlated factor plays a major role in pitted cone and cone chain formation. We propose that pitted cones and cone chains may be related to vertical fractures formed around the UP basin rim from subsidence of infilling basin material.

Analysing Spatial Pattern of the Three Major Fast-Food Chain Restaurants in Johor Bahru Johor, Malaysia

Spatial analysis has proven to be an impactful tool in business for gaining insights and making informed decisions, especially in analysing spatial patterns and finding new locations for starting or expanding the business. However, most fast-food chain restaurants have t yet to optimise the use of GIS, especially in analysing the spatial distribution patterns of their outlets. This study aims to analyse the spatial distribution patterns of three major fast-food chains in Johor Bahru, Malaysia: KFC, McDonald’s, and Burger King. Four spatial analyses were performed: network analysis, buffer analysis, heatmap analysis, and spatial distribution pattern analysis using the nearest neighbour technique. The results showed different spatial distribution patterns for these three major fast-food chains. KFC is more spatially randomly distributed, concentrated in downtown Johor Bahru, while McDonald’s is spatially scattered throughout the Johor Bahru district and the Skudai area. Burger King has too few outlets and is spatially scattered in the Johor Bahru district. From these results, fast-food franchises can use the information to locate new locations based on their indicators for new outlets.

Distraction descriptor for brainprint authentication modelling using probability-based Incremental Fuzzy-Rough Nearest Neighbour

This paper aims to design distraction descriptor, elicited through the object variation, to refine the granular knowledge incrementally, using the proposed probability-based incremental update strategy in Incremental Fuzzy-Rough Nearest Neighbour (IncFRNN) technique. Most of the brainprint authentication models were tested in well-controlled environments to minimize the influence of ambient disturbance on the EEG signals. These settings significantly contradict the real-world situations. Thus, making use of the distraction is wiser than eliminating it. The proposed probability-based incremental update strategy is benchmarked with the ground truth (actual class) incremental update strategy. Besides, the proposed technique is also benchmarked with First-In-First-Out (FIFO) incremental update strategy in K-Nearest Neighbour (KNN). The experimental results have shown equivalence discriminatory performance in both high distraction and quiet conditions. This has proven that the proposed distraction descriptor is able to utilize the unique EEG response towards ambient distraction to complement person authentication modelling in uncontrolled environment. The proposed probability-based IncFRNN technique has significantly outperformed the KNN technique for both with and without defining the window size threshold. Nevertheless, its performance is slightly worse than the actual class incremental update strategy since the ground truth represents the gold standard. In overall, this study demonstrated a more practical brainprint authentication model with the proposed distraction descriptor and the probability-based incremental update strategy. However, the EEG distraction descriptor may vary due to intersession variability. Future research may focus on the intersession variability to enhance the robustness of the brainprint authentication model.

Combining Landsat time series and GEDI data for improved characterization of fuel types and canopy metrics in wildfire simulation

Wildfires in the Chornobyl Exclusion Zone (CEZ) and other radioactively contaminated areas threaten human health and well-being with the potential to resuspend radionuclides. Wildfire behavior simulation is a necessary tool to examine the efficiency of fuel treatments in the CEZ, but it requires systematically updated maps of fuel types and canopy metrics. The objective of this study was to demonstrate an effective approach for mapping fuel types, canopy height (CH), and canopy cover (CC) in territories contaminated by radionuclides using Landsat time series (LTS) and Global Ecosystem Dynamics Investigation (GEDI) LiDAR observations. We combined LTS and GEDI data to map fuel types and canopy metrics used in wildfire simulations within the CEZ. Our classification model showed an adequate overall accuracy (75%) in mapping land covers and associated fuel types. The phenology metrics extracted from LTS reliably distinguished spectrally similar vegetation types (such as grasslands and croplands) which exhibit different flammability through the year. We also predicted a suite of relative heights metrics and CC at Landsat 30-m pixel level (R2 = 0.23–0.26) using the nearest neighbor technique. The imputed maps adequately captured the dynamics of CH and CC in the CEZ after recent large wildfires occurred in 2015, 2020, and 2022. Thus, we illustrate a LTS processing approach to produce wall-to-wall maps of canopy characteristics that are important for wildfire simulations. We conclude that continuous updating of land cover and canopy fuel data is crucial to ensure relevant fire management of radioactively contaminated landscapes and support local decision-making.

The Development of a Consensus Machine Learning Model for Hurricane Rapid Intensification Forecasts with Hurricane Weather Research and Forecasting (HWRF) Data

Abstract This study focused on developing a consensus machine learning (CML) model for tropical cyclone (TC) intensity-change forecasting, especially for rapid intensification (RI). This CML model was built upon selected classical machine learning models with the input data extracted from a high-resolution hurricane model, the Hurricane Weather Research and Forecasting (HWRF) system. The input data contained 21 or 34 RI-related predictors extracted from the 2018 version of HWRF (H218). This study found that TC inner-core predictors can be critical for improving RI predictions, especially the inner-core relative humidity. Moreover, this study emphasized the importance of performing resampling on an imbalanced input dataset. Edited nearest-neighbor and synthetic minority oversampling techniques improved the probability of detection (POD) by ∼10% for the RI class. This study also showed that the CML model has satisfactory performance on RI predictions compared to the operational models. CML reached 56% POD and 46% false alarm ratio (FAR), while the operational models had only 10%–30% POD but 50%–60% FAR. The CML performance on the non-RI classes was comparable to the operational models. The results indicated that, with proper and sufficient training data and RI-related predictors, CML has the potential to provide reliable probabilistic RI forecasts during a hurricane season.

A proposed User-Based Approach for eBooks Recommendation Using a Weighted Nearest Neighbor Technique

Large book data stores were beneficial for our support systems but posed significant challenges for useful information retrieval. This issue was resolved by collaboratively filtering data depending on user needs. This study suggested a user-based methodology for recommending eBooks. The selected dataset was pre-processed, and Cross-validation was used to build a user-user similarity matrix. Three nearest neighbor algorithms (KNN Basic, KNN with Means and KNN with ZScore) were used, and weighted KNN was proposed for rating prediction. In this technique, the weight of each user was calculated based on its distance from the intended user. The evaluation process depends on the user-item matrix and user-user matrix for prediction. The proposed recommendation system was tested on the book-crossing dataset, and the results were evaluated using the root mean square error and the mean absolute value of error. The results show that the error rate of the proposed model is the lowest compared to the other methods used, specifically when using the Pearson-Baseline technique. Since the root mean square error is 1.647 and the mean absolute value of errors is 1.253. When using the cosine technique, the root mean square error is 1.742, and the mean absolute value of errors is 1.328.

Logistic Model Tree Forest for Steel Plates Faults Prediction

Fault prediction is a vital task to decrease the costs of equipment maintenance and repair, as well as to improve the quality level of products and production efficiency. Steel plates fault prediction is a significant materials science problem that contributes to avoiding the progress of abnormal events. The goal of this study is to precisely classify the surface defects in stainless steel plates during industrial production. In this paper, a new machine learning approach, entitled logistic model tree (LMT) forest, is proposed since the ensemble of classifiers generally perform better than a single classifier. The proposed method uses the edited nearest neighbor (ENN) technique since the target class distribution in fault prediction problems reveals an imbalanced dataset and the dataset may contain noise. In the experiment that was conducted on a real-world dataset, the LMT forest method demonstrated its superiority over the random forest method in terms of accuracy. Additionally, the presented method achieved higher accuracy (86.655%) than the state-of-the-art methods on the same dataset.

Complementary Architecture of E-Learning Path Personalization and Optimization

Using computing algorithms to generate personalized learning resources to provide the needs and improve the capabilities, preferences, and academic performance of diverse learners is creating preferred learning environments. As more learning resources, strategies and techniques are frequently added to these e-learning systems, input data to personalize the learning path has been growing exponentially making swift responses to learner’s requests difficult. This study proposed a complementary learning path personalization architecture using ant colony (ACO) with nearest neighbour technique and genetic algorithm (GA) to extend the functions of the Spark framework purposely to develop a robust evolutionary computing algorithm. Experimental results indicate that complementing ACO, GA and Spark frameworks improved the generation of personalized learning resources and best-fitted-optimized learning paths. Spark-ACO took less computational time than standalone ACO. Combining ACO and GA improved the likelihood of an ant colony being trapped at a local optimum, and Spark-ACO-GA significantly enhanced the accuracy of the solutions.

Identification of Essential Protein Using Chemical Reaction Optimization and Machine Learning Technique.

For the survival, development, and reproduction of the organism, understanding the working process of the cell, disease study, design drugs, etc. essential protein plays a crucial role. Due to a large number of biological information, computational methods are becoming popular in recent times to identify the essential protein. Many computational methods used machine learning techniques, metaheuristic algorithms, etc. to solve the problem. The problem with these methods is that the essential protein class prediction rate is still low. Many of these methods have not considered the imbalance characteristics of the dataset. In this paper, we have proposed an approach to identify essential proteins using a metaheuristic algorithm named Chemical Reaction Optimization (CRO) and machine learning method. Both topological and biological features are used here. The Saccharomyces cerevisiae (S. cerevisiae) and Escherichia coli (E. coli) datasets are used in the experiment. Topological features are calculated from the PPI network data. Composite features are calculated from the collected features. Synthetic Minority Over-sampling Technique and Edited Nearest Neighbor (SMOTE+ENN) technique is applied to balance the dataset and then the CRO algorithm is applied to achieve the optimal number of features. Our experiment shows that the proposed approach gives better results in both accuracy and f-measure than the existing related methods.

Recommendation System for a Delivery Food Application Based on Number of Orders

With the recent growth in food-delivery applications, creating new recommendation systems tailored to this platform is essential. State-of-the-art restaurant recommendation systems are based on users’ ratings or reviews, with data that are obtained from questionnaires or online platforms such as TripAdvisor, Zomato, Foursquare, or Yield. However, not all users give ratings or reviews after their purchase. This document proposes a recommendation system whose input is the number of orders stored by a real food-delivery application. These data are always available for all food-delivery applications and are stored all the time. Our proposal is based on the nearest-neighbor technique that calculates the client’s preferred restaurants and analyzes other clients with similar buying patterns. In addition, we propose a performance metric that can be used for this specific recommendation system that is based on real restaurant sales. We use a real dataset (available online) to validate our proposal. Based on our experiments, the recommendation system successfully gives only an average of 7.7 options from 187 that are available. We compared our proposal with other state-of-the-art recommendation techniques and obtained a better performance. Our results indicate that it is possible to generate recommendations based on the number of orders, making the use of a restaurant-recommendation system feasible in a real food-delivery application.

Deep Transfer Learning Based Parkinson’s Disease Detection Using Optimized Feature Selection

Parkinson’s disease (PD) is one of the chronic neurological diseases whose progression is slow and symptoms have similarities with other diseases. Early detection and diagnosis of PD is crucial to prescribe proper treatment for patient’s productive and healthy lives. The disease’s symptoms are characterized by tremors, muscle rigidity, slowness in movements, balancing along with other psychiatric symptoms. The dynamics of handwritten records served as one of the dominant mechanisms which support PD detection and assessment. Several machine learning methods have been investigated for the early detection of this disease. But most of these handcrafted feature extraction techniques predominantly suffer from low performance accuracy issues. This cannot be tolerable for dealing with detection of such a chronic ailment. To this end, an efficient deep learning model is proposed which can assist to have early detection of Parkinson’s disease. The significant contribution of the proposed model is to select the most optimum features which have the effect of getting the high-performance accuracies. The feature optimization is done through genetic algorithm wherein <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$K$ </tex-math></inline-formula> -Nearest Neighbour technique. The proposed novel model results into detection accuracy higher than 95%, precision of 98%, area under curve of 0.90 with a loss of 0.12 only. The performance of proposed model is compared with some state-of-the-art machine learning and deep learning-based PD detection approaches to demonstrate the better detection ability of our model.