High Number Of Dimensions Research Articles

Large-scale IoT attack detection scheme based on LightGBM and feature selection using an improved salp swarm algorithm

Due to the swift advancement of the Internet of Things (IoT), there has been a significant surge in the quantity of interconnected IoT devices that send and exchange vital data across the network. Nevertheless, the frequency of attacks on the Internet of Things is steadily rising, posing a persistent risk to the security and privacy of IoT data. Therefore, it is crucial to develop a highly efficient method for detecting cyber threats on the Internet of Things. Nevertheless, several current network attack detection schemes encounter issues such as insufficient detection accuracy, the curse of dimensionality due to excessively high data dimensions, and the sluggish efficiency of complex models. Employing metaheuristic algorithms for feature selection in network data represents an effective strategy among the myriad of solutions. This study introduces a more comprehensive metaheuristic algorithm called GQBWSSA, which is an enhanced version of the Salp Swarm Algorithm with several strategy improvements. Utilizing this algorithm, a threshold voting-based feature selection framework is designed to obtain an optimized set of features. This procedure efficiently decreases the number of dimensions in the data, hence preventing the negative effects of having a high number of dimensions and effectively extracting the most significant and crucial information. Subsequently, the extracted feature data is combined with the LightGBM algorithm to form a lightweight and efficient ensemble learning scheme for IoT attack detection. The proposed enhanced metaheuristic algorithm has superior performance in feature selection compared to the recent metaheuristic algorithms, as evidenced by the experimental evaluation conducted using the NSLKDD and CICIoT2023 datasets. Compared to current popular ensemble learning solutions, the proposed overall solution exhibits excellent performance on multiple key indicators, including accuracy, precision, as well as training and detection time. Especially on the large-scale dataset CICIoT2023, the proposed scheme achieves an accuracy rate of 99.70% in binary classification and 99.41% in multi classification.

Unlocking Online Insights: LSTM Exploration and Transfer Learning Prospects

Machine learning algorithms can improve the time series data analysis as compared to the traditional methods such as moving averages or auto-regressive approaches. This advancement has helped to unlock several challenging problems since machine learning not only helps to forecast the overall trend of the data, but it also helps to keep the historical track of changes in factors, influencing this trend. These predictions play a pivotal role in almost all areas of research where the observations are time dependent, such as problems ranging from challenges of finance to public health, environmental and climate change challenges. A key challenge of these domains is the higher number of attributes and predictors since managing and manipulating data from many attributes is itself a significant challenge for future forecasting. Addressing these challenges is possible with Recursive Long Short-Term Memory models. The application of such models is crucial, and their efficacy is further amplified when considering transfer learning. During this research, a detailed and comprehensive description of such models is addressed. Practical application is illustrated through an example, emphasizing that these models, when transferred to complex and large datasets using transfer learning, hold great promise.

Graphical representation of data prediction potential: correlation graphs and correlation chains

The correlation of the set of attributes is a crucial statistical value for the measuring of prediction potential present in a dataset. The correlation coefficient, which measures the correlation between the values of two attributes, can be used in order to measure the prediction potential between two-element subsets of a dataset containing a high number of attributes. In this way two common summary visualizations of prediction potential in datasets are formed—correlation matrices and correlation heatmaps. Both of these visualizations are focused on the presentation of correlation between pair of attributes but not much more regarding the context of correlations in the dataset. The main objective of this article is the design and implementation of graphical models usable in a visual representation of data prediction potential—correlation graphs and correlation chains—which emphasize the pseudo-transitivity of prediction potential in a dataset.

Patterns and correlates of old-age social exclusion in the Balkan states

Social exclusion (SE), or the separation of individuals and groups from mainstream society, is associated with poor health and wellbeing, yet a substantial number of older people are socially excluded. There is increasing agreement that SE is multidimensional, comprising among others social relations, material resources, and/or civic participation. However, measuring SE is still challenging as exclusion may occur in more than one dimension, whereas its sum does not reflect the content of SE. To account for these challenges, this study provides a typology of SE and describes how SE types differ from each other in terms of severity and risk factors. We concentrate on Balkan states, which are among the European countries with the highest prevalence of SE. Data come from the European Quality of Life Survey (N = 3030, age 50 +). Latent Class Analysis revealed four SE types: low SE risk (50%), material exclusion (23%), material and social exclusion (4%), and multidimensional exclusion (23%). A higher number of dimensions from which a person is excluded are associated with more severe outcomes. Multinomial regression further revealed that lower levels of education, lower subjective health, and lower social trust increase the risks of any SE type. Younger age, unemployment, and not having a partner are associated with specific SE types. This study is in line with the limited evidence that different types of SE exist. Policies designed to reduce SE should take account of the different SE types and specific associated risk factors in order to enhance the impact of interventions to reduce social exclusion.

Naïve Bayes dan Support Vector Machine Berbasis PSO untuk Seleksi Fitur pada Sentiment Analysis

Sentiment analysis is a process that aims to determine the content of the dataset in the form of positive, negative and neutral text. Currently the opinion of the general public is an important source of decision making. Social media is a place to express public opinion on an object, problem or event. Such as the government's policy regarding the relocation of the capital city of Indonesia, which was originally in Jakarta to Kalimantan, did not escape the attention of the public, especially Twitter users. One of the problems in sentiment analysis is the high number of attributes and dimensions in the dataset. In this study, sentiment analysis was carried out on the relocation of the national capital using the Naïve Bayes method, and the Support Vector Machine based on Particle Swarm Optimization (PSO). The advantages of the Support Vector Machine are High dimensional space and Vector document space. Feature selection greatly affects the performance of the classification, the use of PSO as feature selection to improve accuracy. The results of this study obtained the best accuracy value of 96.45% and the AUC value of 0.920 from the application of PSO on the Support Vector Machine. This result has increased when compared to the experimental results using Naïve Bayes and Support Vector Machine without PSO. The application of PSO on the Support Vector Machine is proven to get better accuracy results in predicting sentiment analysis on the dataset of moving the State Capital.

Strategy selection in decisions from givens: Deciding at a glance?

People deciding between alternatives have at their disposal a toolbox containing both compensatory strategies, which take into account all available attributes of those alternatives, and noncompensatory strategies, which consider only some of the attributes. It is commonly assumed that noncompensatory strategies play only a minor role in decisions from givens, where attribute information is openly presented, because all attributes can be processed automatically “at a glance.” Based on a literature review, however, I establish that previous studies on strategy selection in decisions from givens have yielded highly heterogeneous findings, including evidence of widespread use of noncompensatory strategies. Drawing on insights from visual attention research on subitizing, I argue that this heterogeneity might be due to differences across studies in the number of attributes and in whether the same or different symbols are used to represent high/low attribute values across attributes. I tested the impact of these factors in two experiments with decisions from givens in which both the number of attributes shown for each alternative and the coding of attribute values was manipulated. An analysis of participants’ strategy use with a Bayesian multimethod approach (taking into account both decisions and response-time patterns) showed that a noncompensatory strategy was more frequently selected in conditions with a higher number of attributes; the type of attribute coding scheme did not affect strategy selection. Using a compensatory strategy in the conditions with eight (vs. four) attributes was associated with rather long response times and a high rate of strategy execution errors. The results suggest that decisions from givens can incur cognitive costs that prohibit reliance on automatic compensatory decision making and that can favor the adaptive selection of a noncompensatory strategy.

Optimized time reduction models applied to power and energy systems planning – Comparison with existing methods

The paper proposes a strategy for the time horizon reduction in power and energy studies. The method denoted Optimized Weighted Time Slices is compared with conventional approaches based on representative days that rely on unsupervised and supervised clustering as well as different strategies to reconstruct the problems. Those reference methods suffer from a lack of scalability when high numbers of dimensions are considered and their outputs strongly depend on the starting point in the partitioning process. The proposed strategy is based on a hierarchical clustering coupled with a least square minimization. The originality of the approach is that it works on individual time slices rather than on representative periods. Those representative time samples are furtherly optimized considering fitting criteria with the input time series thanks to a linearization of the duration curves. All the time modelling methods are tested on both a simple energy hub at a building scale (i.e. load, solar storage) and on a 33-buses distribution network with storage. The methods performances are assessed while comparing the results of the systems operation over the reduced time horizons with the outputs from full yearly simulations. In particular, complex objective functions are considered for the systems operation, as it is shown that they impact the accuracy of the time reduction as much as the systems complexity itself. The proposed strategy displays smaller errors (1%–5% more accuracy) than the reference methods, is much more scalable (>10 times faster), and systematically returns the same outputs.

The application of Uniform Manifold Approximation and Projection (UMAP) for unconstrained ordination and classification of biological indicators in aquatic ecology

The analysis of community structure in studies of freshwater ecology often requires the application of dimensionality reduction to process multivariate data. A high number of dimensions (number of taxa/environmental parameters × number of samples), nonlinear relationships, outliers, and high variability usually hinder the visualization and interpretation of multivariate datasets. Here, we proposed a new statistical design using Uniform Manifold Approximation and Projection (UMAP), and community partitioning using Louvain algorithms, to ordinate and classify the structure of aquatic biota in two-dimensional space. We present this approach with a demonstration of five previously published datasets for diatoms, macrophytes, chironomids (larval and subfossil), and fish. Principal Component Analysis (PCA) and Ward's clustering were also used to assess the comparability of the UMAP approach compared to traditional approaches for ordination and classification. The ordination of sampling sites in 2-dimensional space showed a much denser, and easier to interpret, grouping using the UMAP approach in comparison to PCA. The classification of community structure using the Louvain algorithm in UMAP ordinal space showed a high classification strength for data with a high number of dimensions than the cluster patterns obtained with the use of a Ward's algorithm in PCA. Environmental gradients, presented via heat maps, were overlayed with the ordination patterns of aquatic communities, confirming that the ordinations obtained by UMAP were ecologically meaningful. This is the first study that has applied a UMAP approach with classification using Louvain algorithms on ecological datasets. We show that the performance of local and global structures, as well as the number of clusters determined by the algorithm, make this approach more powerful than traditional approaches.

Cosmological parameter estimation and inference using deep summaries

The ability to obtain reliable point estimates of model parameters is of crucial importance in many fields of physics. This is often a difficult task given that the observed data can have a very high number of dimensions. In order to address this problem, we propose a novel approach to construct parameter estimators with a quantifiable bias using an order expansion of highly compressed deep summary statistics of the observed data. These summary statistics are learned automatically using an information maximising loss. Given an observation, we further show how one can use the constructed estimators to obtain approximate Bayes computation (ABC) posterior estimates and their corresponding uncertainties that can be used for parameter inference using Gaussian process regression even if the likelihood is not tractable. We validate our method with an application to the problem of cosmological parameter inference of weak lensing mass maps. We show in that case that the constructed estimators are unbiased and have an almost optimal variance, while the posterior distribution obtained with the Gaussian process regression is close to the true posterior and performs better or equally well than comparable methods.

Permission-based Android malware analysis by using dimension reduction with PCA and LDA

The expansion of the Android operating system increases the interest of malware developers in this field. With the develop malicious software, the material, and moral harms are given to the users in many ways such as stealing personal data and decreasing device performance. Therefore, the need for systems that detect malware with high accuracy is increasing day by day. In this study, it is aimed to detect malware with a static analysis technique based on machine learning. Because of the limited system resources of mobile devices, principal component analysis and linear discriminant analysis, which are frequently used in machine learning problems with a high number of attributes, are applied for Android malware detection. When the results are examined, it is observed that the dimension reduction techniques have a positive effect on classification performance in general.

A scalable software solution for anonymizing high-dimensional biomedical data.

BackgroundData anonymization is an important building block for ensuring privacy and fosters the reuse of data. However, transforming the data in a way that preserves the privacy of subjects while maintaining a high degree of data quality is challenging and particularly difficult when processing complex datasets that contain a high number of attributes. In this article we present how we extended the open source software ARX to improve its support for high-dimensional, biomedical datasets.FindingsFor improving ARX's capability to find optimal transformations when processing high-dimensional data, we implement 2 novel search algorithms. The first is a greedy top-down approach and is oriented on a formally implemented bottom-up search. The second is based on a genetic algorithm. We evaluated the algorithms with different datasets, transformation methods, and privacy models. The novel algorithms mostly outperformed the previously implemented bottom-up search. In addition, we extended the GUI to provide a high degree of usability and performance when working with high-dimensional datasets.ConclusionWith our additions we have significantly enhanced ARX's ability to handle high-dimensional data in terms of processing performance as well as usability and thus can further facilitate data sharing.

Latent Class Model with Heterogeneous Decision Rule for Identification of Factors to the Choice of Drivers’ Seat Belt Use

The choice of not buckling a seat belt has resulted in a high number of deaths worldwide. Although extensive studies have been done to identify factors of seat belt use, most of those studies have ignored the presence of heterogeneity across vehicle occupants. Not accounting for heterogeneity might result in a bias in model outputs. One of the main approaches to capture random heterogeneity is the employment of the latent class (LC) model by means of a discrete distribution. In a standard LC model, the heterogeneity across observations is considered while assuming the homogeneous utility maximization for decision rules. However, that notion ignores the heterogeneity in the decision rule across individual drivers. In other words, while some drivers make a choice of buckling up with some characteristics, others might ignore those factors while making a choice. Those differences could be accommodated for by allowing class allocation to vary based on various socio-economic characteristics and by constraining some of those rules at zeroes across some of the classes. Thus, in this study, in addition to accounting for heterogeneity across individual drivers, we accounted for heterogeneity in the decision rule by varying the parameters for class allocation. Our results showed that the assignment of various observations to classes is a function of factors such as vehicle type, roadway classification, and vehicle license registration. Additionally, the results showed that a minor consideration of the heterogeneous decision rule resulted in a minor gain in model fits, as well as changes in significance and magnitude of the parameter estimates. All of this was despite the challenges of fully identifying exact attributes for class allocation due to the inclusion of high number of attributes. The findings of this study have important implications for the use of an LC model to account for not only the taste heterogeneity but also heterogeneity across the decision rule to enhance model fit and to expand our understanding about the unbiased point estimates of parameters.

A Novel Density-based Technique for Outlier Detection of High Dimensional Data Utilizing Full Feature Space

Recently, anomaly detection has acquired a realistic response from data mining scientists as a graph of its reputation has increased smoothly in various practical domains like product marketing, fraud detection, medical diagnosis, fault detection and so many other fields. High dimensional data subjected to outlier detection poses exceptional challenges for data mining experts and it is because of natural problems of the curse of dimensionality and resemblance of distant and adjoining points. Traditional algorithms and techniques were experimented on full feature space regarding outlier detection. Customary methodologies concentrate largely on low dimensional data and hence show ineffectiveness while discovering anomalies in a data set comprised of a high number of dimensions. It becomes a very difficult and tiresome job to dig out anomalies present in high dimensional data set when all subsets of projections need to be explored. All data points in high dimensional data behave like similar observations because of its intrinsic feature i.e., the distance between observations approaches to zero as the number of dimensions extends towards infinity. This research work proposes a novel technique that explores deviation among all data points and embeds its findings inside well established density-based techniques. This is a state of art technique as it gives a new breadth of research towards resolving inherent problems of high dimensional data where outliers reside within clusters having different densities. A high dimensional dataset from UCI Machine Learning Repository is chosen to test the proposed technique and then its results are compared with that of density-based techniques to evaluate its efficiency.

Extracting Cell Patterns From High-Dimensional Radio Network Performance Datasets Using Self-Organizing Maps and K-Means Clustering

Mobile Radio Networks produces many of Operations, Administration, and Maintenance (OAM) data used by operators for network operational assurance. These data include multiple and diverse performance measurements and indicators that characterize the behavior of the radio cells. Being able to properly cluster the apparently dissimilar behaviors exhibited by a large number of individual cells into a reduced set of prototype patterns constitutes a valuable tool to support multiple processes such as cell configuration optimization or fault performance root cause analysis. While powerful clustering methods such as Self Organized Maps (SOM) exist, there is practically no literature showing the applicability of these methods of OAM datasets with a high number of attributes (>20) collected from live network deployments. Moreover, the applicability of the clustering methods does not come free of open questions since, for instance, when using SOM there is no explicitly obtained information about clusters after the SOM training in the underlying data, so the k-means technique for grouping SOM units has to be applied afterward. In this context, this paper describes a methodology to cluster radio cells based on a combination of SOM and K-means methods. The methodology is applied to extract cell patterns of the characterization of the long-term behavior (15 days' observation period) and short-term behavior (hourly observation periods) of mobile cells. OAM datasets collected from a live 4G/LTE network deployed in a major European city are used in the analysis.

Improved Deep Distributed Light Field Coding

Light fields enable increasing the degree of realism and immersion of visual experience by capturing a scene with a higher number of dimensions than conventional 2D imaging. On another side, higher dimensionality entails significant storage and transmission overhead compared to traditional video. Conventional coding schemes achieve high coding gains by employing an asymmetric codec design, where the encoder is significantly more complex than the decoder. However, in the case of light fields, the communication and processing among different cameras could be expensive, and the possibility of trading the complexity between the encoder and the decoder becomes a desirable feature. We leverage the distributed source coding paradigm to effectively reduce the encoder’s complexity at the cost of increased computation at the decoder side. Specifically, we train two deep neural networks to improve the two most critical parts of a distributed source coding scheme: the prediction of side information and the estimation of the uncertainty in the prediction. Experiments show considerable BD-rate gains, above 59% over HEVC-Intra and 17.45% over our previous method DLFC-I.

Stratifying patients using fast multiple kernel learning framework: case studies of Alzheimer\u2019s disease and cancers

BackgroundPredictive patient stratification is greatly emerging, because it allows us to prospectively identify which patients will benefit from what interventions before their condition worsens. In the biomedical research, a number of stratification methods have been successfully applied and have assisted treatment process. Because of heterogeneity and complexity of medical data, it is very challenging to integrate them and make use of them in practical clinic. There are two major challenges of data integration. Firstly, since the biomedical data has a high number of dimensions, combining multiple data leads to the hard problem of vast dimensional space handling. The computation is enormously complex and time-consuming. Secondly, the disparity of different data types causes another critical problem in machine learning for biomedical data. It has a great need to develop an efficient machine learning framework to handle the challenges.MethodsIn this paper, we propose a fast-multiple kernel learning framework, referred to as fMKL-DR, that optimise equations to calculate matrix chain multiplication and reduce dimensions in data space. We applied our framework to two case studies, Alzheimer’s disease (AD) patient stratification and cancer patient stratification. We performed several comparative evaluations on various biomedical datasets.ResultsIn the case study of AD patients, we enhanced significantly the multiple-ROIs approach based on MRI image data. The method could successfully classify not only AD patients and non-AD patients but also different phases of AD patients with AUC close to 1. In the case study of cancer patients, the framework was applied to six types of cancers, i.e., glioblastoma multiforme cancer, ovarian cancer, lung cancer, breast cancer, kidney cancer, and liver cancer. We efficiently integrated gene expression, miRNA expression, and DNA methylation. The results showed that the classification model basing on integrated datasets was much more accurate than classification model basing on the single data type.ConclusionsThe results demonstrated that the fMKL-DR remarkably improves computational cost and accuracy for both AD patient and cancer patient stratification. We optimised the data integration, dimension reduction, and kernel fusion. Our framework has great potential for mining large-scale cohort data and aiding personalised prevention.

Analytical review of clustering techniques and proximity measures

One of the most fundamental approaches to learn and understand from any type of data is by organizing it into meaningful groups (or clusters) and then analyzing them, which is a process known as cluster analysis. During this process of grouping, proximity measures play a significant role in deciding the similarity level of two objects. Moreover, before applying any learning algorithm on a dataset, different aspects related to preprocessing such as dealing with the sparsity of data, leveraging the correlation among features and normalizing the scales of different features are required to be considered. In this study, various proximity measures have been discussed and analyzed from the aforementioned aspects. In addition, a theoretical procedure for selecting a proximity measure for clustering purpose is proposed. This procedure can also be used in the process of designing a new proximity measure. Second, clustering algorithms of different categories have been overviewed and experimentally compared for various datasets of different domains. The datasets have been chosen in such a way that they range from a very low number of dimensions to a very high number of dimensions. Finally, the effect of using different proximity measures is analyzed in partitional and hierarchical clustering techniques based on experiments.

Exploring Different Paradigms to Extract Proper Implications From High Dimensional Formal Contexts

Formal Concept Analysis (FCA) is an applied mathematical technique for data analysis, in which the relations between objects and attributes are identified. It introduces the notion of concepts and their hierarchical structure, from which we can obtain a set of implications between attributes that characterize a knowledge domain. The volume of information to be processed makes the use of FCA difficult in domains with a high number of dimensions, creating a demand for new solutions and algorithms for FCA applications. This article explores different approaches to extract proper implications from high dimensional contexts based on constraints to obtain the set of implications rules. We propose algorithms that use a data structure called Binary Decision Diagram (BDD) to represent the formal context, which reduces its size and, due to this, operates more efficiently. We also propose a heuristic to obtain proper implications by reducing the unnecessary generation of premises. In addition, we implemented a parallel computing model for generating and obtaining different implications. To analyze the proposed algorithms, we used different synthetic contexts with a varying number of objects, attributes, and density. The results obtained presented speed gains of up to 22 times when compared to the solutions proposed in the literature such as Impec and PropIm.

Local Neighborhood-based Outlier Detection of High Dimensional Data using different Proximity Functions

In recent times, dimension size has posed more challenges as compared to data size. The serious concern of high dimensional data is the curse of dimensionality and has ultimately caught the attention of data miners. Anomaly detection based on local neighborhood like local outlier factor has been admitted as state of art approach but fails when operated on the high number of dimensions for the reason mentioned above. In this paper, we determine the effects of different distance functions on an unlabeled dataset while digging outliers through the density-based approach. Further, we also explore findings regarding runtime and outlier score when dimension size and number of nearest neighbor points (min_pts) are varied. This analytic research is also very appropriate and applicable in the domain of big data and data science as well.

An accelerated gradient algorithm for well control optimization

The simulation-assisted optimization of injection and production control settings can be considered as a way for improved management of hydrocarbon reservoirs. Although well control optimization problems are expected to have a relatively smooth landscape, they suffer from having a high number of dimensions. As a result, gradient-based algorithms are typically the preferred choice over derivative-free algorithms. Since the calculation of the exact gradient is not always feasible, recent studies into gradient-based algorithms have focused on ameliorating the accuracy of gradient approximation methods. Less attention has been paid to the framework in which these gradient approximations are utilized. The most common framework is the steepest descent with a backtracking line-search. This study aims to provide an improved framework using an adaptive moment estimation technique. The proposed method is compared to the steepest descent framework in two case studies of increasing complexity. In both frameworks, the gradient is approximated by the simultaneous perturbation stochastic approximation. The first case study is a two-dimensional heterogeneous reservoir model produced under water-flooding. The second case study investigates the injection and production settings of the three-dimensional benchmark case, the Brugge model, whilst incorporating geological uncertainty. The proposed framework showed improvements of up to 91% in convergence speeds and up to 5% in optimal value for the first case study. In addition, the proposed framework showed an improvement of up to 81% in convergence speeds and up to 2% in optimal value for the second case study over the steepest descent framework. For both case studies, the effect of the gradient approximation accuracy (perturbation number) was also investigated. The results indicated that the proposed algorithm is less sensitive to the gradient approximation accuracy than the steepest descent framework. In addition, this study investigated the effect of two popular bound constraint handling techniques in both frameworks. The results indicated that the projection method outperforms the logarithmic transformation method when applied to production optimization problem.