Curse Of Dimensionality Problem Research Articles

Temporal Sequential-Artificial Neural Network Enhancements for Improved Smart Lighting Control

Several previous studies have proposed a temporal sequential-artificial neural network (TS-ANN) to convert PIR Sensor movement data into presence data and improve the performance of smart lighting control. However, such a temporal-sequential forecasting concept has a curse of dimensionality problem. This study aims to proposes the application of principal component analysis with TS-ANN (PCA-TS-ANN) as an intelligent method for controlling smart lighting with low dimensions. We have primary data directly from a smart lighting implementation that utilizes PIR sensors. We apply cross-correlation to the original dataset to find the optimum time step. Then we discover the optimum TS-ANN based on selected tuning parameter values through PCC. We then design and compare scenarios involving the combination of TS-ANN and PCA. Finally, we evaluate these scenarios using the metrics Accuracy, Precision, Recall, F1− Score, and Delay. The results of this study are the PCA-TS-ANN model with Accuracy, Precision, Recall, and F1−Score value of 0.9993, 0.9997, 0.9994, and 0.9996 respectively. The PCA method reduces the TS-ANN features from 1200 features to 36 features. The model size has also decreased from 3534kB to 807kB. Our model has a simpler complexity with TS-ANN that the µ ± σ Delay is 0.27±0.06 ms versus 0.34±0.11 ms.

P2S distance induced locally conjugated orthogonal subspace learning for feature extraction

When performing data classification tasks, it often occurs to them the curse of dimensionality problem. To address the issue, a manifold learning method termed locally conjugated orthogonal subspace (LCOS) is put forward for dimensionality reduction or feature extraction in this paper. Note that point to feature space (P2S) distance contributes to mining local geometry information, both a local margin characterizing data apartness and a locally conjugated orthogonal constraint beneficial to removing data redundancy are well studied from the P2S distance metric. They are all exploited to model the proposed LCOS. Then, a low dimensional subspace can be explored by maximizing the P2S distance induced local margin under the constraint. Compared with some other related dimensionality reduction methods, experimental results on benchmark face and object data sets validate the performance of the proposed method.

Nonparametric Estimation for High-Dimensional Space Models Based on a Deep Neural Network

With high dimensionality and dependence in spatial data, traditional parametric methods suffer from the curse of dimensionality problem. The theoretical properties of deep neural network estimation methods for high-dimensional spatial models with dependence and heterogeneity have been investigated only in a few studies. In this paper, we propose a deep neural network with a ReLU activation function to estimate unknown trend components, considering both spatial dependence and heterogeneity. We prove the compatibility of the estimated components under spatial dependence conditions and provide an upper bound for the mean squared error (MSE). Simulations and empirical studies demonstrate that the convergence speed of neural network methods is significantly better than that of local linear methods.

A Hybrid Metaheuristics based technique for Mutation Based Disease Classification

Due to recent advancements in computational biology, DNA microarray technology has evolved as a useful tool in the detection of mutation among various complex diseases like cancer. The availability of thousands of microarray datasets makes this field an active area of research. Early cancer detection can reduce the mortality rate and the treatment cost. Cancer classification is a process to provide a detailed overview of the disease microenvironment for better diagnosis. However, the gene microarray datasets suffer from a curse of dimensionality problems also the classification models are prone to be overfitted due to small sample size and large feature space. To address these issues, the authors have proposed an Improved Binary Competitive Swarm Optimization Whale Optimization Algorithm (IBCSOWOA) for cancer classification, in which IBCSO has been employed to reduce the informative gene subset originated from using minimum redundancy maximum relevance (mRMR) as filter method. The IBCSOWOA technique has been tested on an artificial neural network (ANN) model and the whale optimization algorithm (WOA) is used for parameter tuning of the model. The performance of the proposed IBCSOWOA is tested on six different mutation-based microarray datasets and compared with existing disease prediction methods. The experimental results indicate the superiority of the proposed technique over the existing nature-inspired methods in terms of optimal feature subset, classification accuracy, and convergence rate. The proposed technique has illustrated above 98% accuracy in all six datasets with the highest accuracy of 99.45% in the Lung cancer dataset.

RECENT DEVELOPMENTS IN SOFT COMPUTING BASED TECHNIQUES FOR FEATURE SELECTION AND DISEASE CLASSIFICATION

Computational prediction of diseases is vital in medical research that contributes to computer-aided diagnostics and helps doctors and medical practitioners in critical decision-making for various diseases such as bacterial and viral kinds of disease, including COVID-19 of the current pandemic situation. Feature selection techniques function as a preprocessing phase for classification and prediction algorithms. For disease prediction, these features may be the patient’s clinical profiles or genomic features such as gene expression profiles from microarray and read counts from RNA-Seq. The performance of a classifier depends primarily on the selected features. In addition, genomic features are too large in numbers, resulting in the curse of dimensionality problem. In the last few years, several feature selection algorithms have been developed to overcome the existing problems to get rid of eliminating chronic diseases, such as various cancers, Zika virus, Ebola virus, and the COVID-19 pandemic. In this review article, we systematically associate soft computing-based approaches for feature selection and disease prediction by applying three data types: patients’ clinical profiles, microarray gene expression profiles, and RNA-Seq sample profiles. According to related work, when the discussion took place, the percentage of medical data types highlighted through pictorial representation and the respective ratio of percentages mentioned were 52%, 27%, 9% and 12% for clinical symptoms, gene expression, MRI-Image and other data types such as signal or text-based utilized, respectively. We also highlight the significant challenges and future directions in this research domain.

A Class Activation Map-Based Interpretable Transfer Learning Model for Automated Detection of ADHD from fMRI Data.

Automatic detection of Attention Deficit Hyperactivity Disorder (ADHD) based on the functional Magnetic Resonance Imaging (fMRI) through Deep Learning (DL) is becoming a quite useful methodology due to the curse of-dimensionality problem of the data is solved. Also, this method proposes an invasive and robust solution to the variances in data acquisition and class distribution imbalances. In this paper, a transfer learning approach, specifically ResNet-50 type pre-trained 2D-Convolutional Neural Network (CNN) was used to automatically classify ADHD and healthy children. The results demonstrated that ResNet-50 architecture with 10-k cross-validation (CV) achieves an overall classification accuracy of 93.45%. The interpretation of the results was done via the Class Activation Map (CAM) analysis which showed that children with ADHD differed from controls in a wide range of brain areas including frontal, parietal and temporal lobes.

Chicken Swarm-Based Feature Subset Selection with Optimal Machine Learning Enabled Data Mining Approach

Data mining (DM) involves the process of identifying patterns, correlation, and anomalies existing in massive datasets. The applicability of DM includes several areas such as education, healthcare, business, and finance. Educational Data Mining (EDM) is an interdisciplinary domain which focuses on the applicability of DM, machine learning (ML), and statistical approaches for pattern recognition in massive quantities of educational data. This type of data suffers from the curse of dimensionality problems. Thus, feature selection (FS) approaches become essential. This study designs a Feature Subset Selection with an optimal machine learning model for Educational Data Mining (FSSML-EDM). The proposed method involves three major processes. At the initial stage, the presented FSSML-EDM model uses the Chicken Swarm Optimization-based Feature Selection (CSO-FS) technique for electing feature subsets. Next, an extreme learning machine (ELM) classifier is employed for the classification of educational data. Finally, the Artificial Hummingbird (AHB) algorithm is utilized for adjusting the parameters involved in the ELM model. The performance study revealed that FSSML-EDM model achieves better results compared with other models under several dimensions.

A Machine Learning-Based Reliability Evaluation Model for Integrated Power-Gas Systems

This paper proposes a machine learning method for the reliability evaluation of integrated power-gas systems (IPGS) under the uncertain component failure probability distributions. The Random Forest (RF) method is designed to select important features to solve the insufficient quantity of data and the curse of dimensionality problems. The Extreme Gradient Boosting (XGBoost) regression algorithm is developed to quantify the relationship between the uncertain parameters and reliability metrics. Moreover, a ten-fold cross-validation method is employed to further improve the accuracy of the regression model. Simulation results on three test systems show that the proposed method can achieve high accuracy for the reliability evaluation.

Random Forests and the measurement of super-efficiency in the context of Free Disposal Hull

In the technical efficiency evaluation area, it may happen that many observations obtain a similar relative technical efficiency status, making it difficult to discriminate between them. The determination of super-efficiency has been a way of solving this problem by providing a method to differentiate between the performance of observations. Despite the existence of some approaches dealing with the notion of super-efficiency in the literature, there have been few attempts to address this problem from the standpoint of machine learning techniques. In this paper, we fill this gap by adapting Random Forest to determine super-efficiency in the context of the Free Disposal Hull (FDH) technique. The new super-efficiency approach is robust to resampling on inputs and data. Additionally, we show how the new approach could be a possible solution for dealing with the curse of dimensionality problem; typically associated with FDH. Furthermore, exploiting the adaptation of Random Forest, a new method for assessing the importance of input variables is introduced. Finally, the advantages of the proposed approach are illustrated through a real example.

Student Performance Prediction Using A Cascaded Bi-level Feature Selection Approach

Features in educational data are ambiguous which leads to noisy features and curse of dimensionality problems. These problems are solved via feature selection. There are existing models for features selection. These models were created using either a single-level embedded, wrapperbased or filter-based methods. However single-level filter-based methods ignore feature dependencies and ignore the interaction with the classifier. The embedded and wrapper based feature selection methods interact with the classifier, but they can only select the optimal subset for a particular classifier. So their selected features may be worse for other classifiers. Hence this research proposes a robust Cascade Bi-Level (CBL) feature selection technique for student performance prediction that will minimize the limitations of using a single-level technique. The proposed CBL feature selection technique consists of the Relief technique at first-level and the Particle Swarm Optimization (PSO) at the second-level. The proposed technique was evaluated using the UCI student performance dataset. In comparison with the performance of the single-level feature selection technique the proposed technique achieved an accuracy of 94.94% which was better than the values achieved by the single-level PSO with an accuracy of 93.67% for the binary classification task. These results show that CBL can effectively predict student performance.

Developing an attack detection framework for wireless sensor network‐based healthcare applications using hybrid convolutional neural network

AbstractAttack detection is the major issue in healthcare‐based wireless sensor networks (H‐WSNs). Due to their low processing speed, very low storage space, poor attack detection rate, longer deployment time, poor communication range, and reduced energy, H‐WSNs are subjected to difficult implementation and have their own limitations. To tackle these issues, we have presented a hybrid deep learning model using convolutional neural network and long short term memory (HDMCL) for attack detection in H‐WSN. This research is divided into three steps: preprocessing, dimensionality reduction, and classification (attack detection). At first, the raw input data (patient's health data) is preprocessed using the one‐hot encoding method. Next, the modified Huber independent component analysis based squirrel search algorithm (MHICA‐SSA) effectively reduces the data dimensionality in which the MHICA‐SSA is the amalgamation of both modified Huber independent component analysis and squirrel search algorithm. The novel algorithm designed overcomes the complexities associated with existing techniques such as the curse of dimensionality problem, improves the parameter interpretation, minimizes the time and storage space, minimizes space complexity, and enhances the attack detection accuracy of the convolutional neural network‐based long short‐term memory (CNN‐LSTM). After that, the deep learning CNN‐LSTM model is utilized for normal, black hole, and gray hole attack detection. The NS2.34 network simulator implements the proposed work thereby the proposed work efficiency is validated using different performance measures such as packet delivery ratio, false alarm rate, network lifetime, energy consumption, throughput, attack detection time, and rate. The proposed work performance is evaluated using different existing methods such as fish swarm optimization based particle swarm optimization, intelligent opportunistic routing algorithm, Jensen‐Shannon divergence‐based independent component analysis, and radio frequency identification based WSN. The throughput of the proposed work is increased up to 3.2 times and the network lifetime is increased up to 4 times when compared to the existing techniques.

An efficient parallel indexing structure for multi-dimensional big data using spark

With the increasing daily production of data in recent years, indexing, storing and retrieving huge amounts of data have become a common problem, especially for multi-dimensional big data. Although R-tree has proved to be efficient for indexing multi-dimensional big data, the R-tree suffers from the curse of dimensionality problem. Many researchers continue to use the R-tree in their studies as it is the most famous tree-like structure for indexing multi-dimensional data. However, with increasing numbers of dimensions in multi-dimensional data the performance of R-Tree will decrease. This paper proposes a new indexing structure called Parallel Indexing System Structure based on Spark (ParISSS), which is an efficient system for indexing multi-dimensional big data, to overcome these problems. ParISSS introduces six types of computing nodes, the reception-node is used to insert and index data, the normal-node is used to store indexed data, the resolution-node is used to distribute a reception-index to a normal-node, the representative-node is used to receive queries from the user, and the reply-node and check-node are used to send the results to the user. We also introduced BR*-tree structure to improve the storing and searching processes. We present an extensive experimental evaluation of our system, comparing several indexing systems. The experimental results show that ParISSS outperforms other indexing systems.

Kernel Identification of Non-Linear Systems with General Structure

In the paper we deal with the problem of non-linear dynamic system identification in the presence of random noise. The class of considered systems is relatively general, in the sense that it is not limited to block-oriented structures such as Hammerstein or Wiener models. It is shown that the proposed algorithm can be generalized for two-stage strategy. In step 1 (non-parametric) the system is approximated by multi-dimensional regression functions for a given set of excitations, treated as representative set of points in multi-dimensional space. ‘Curse of dimensionality problem’ is solved by using specific (quantized or periodic) input sequences. Next, in step 2, non-parametric estimates can be plugged into least squares criterion and support model selection and estimation of system parameters. The proposed strategy allows decomposition of the identification problem, which can be of crucial meaning from the numerical point of view. The “estimation points” in step 1 are selected to ensure good task conditioning in step 2. Moreover, non-parametric procedure plays the role of data compression. We discuss the problem of selection of the scale of non-parametric model, and analyze asymptotic properties of the method. Also, the results of simple simulation are presented, to illustrate functioning of the method. Finally, the proposed method is successfully applied in Differential Scanning Calorimeter (DSC) to analyze aging processes in chalcogenide glasses.

Linguistic feature based learning model for fake news detection and classification

Social media is used as a dominant source of news distribution among users. The world's preeminent decisions such as politics are acclaimed by social media to influence users for enclosing users' decisions in their favor. However, the adoption of social media is much needed for awareness but the authenticity of content is an unknown factor in the current scenario. Therefore, this research work finds it imperative to propose a solution to fake news detection and classification. In the case of fake news, content is the prime entity that captures the human mind towards trust for specific news. Therefore, a linguistic model is proposed to find out the properties of content that will generate language-driven features. This linguistic model extracts syntactic, grammatical, sentimental, and readability features of particular news. Language driven model requires an approach to handle time-consuming and handcrafted features problems in order to deal with the curse of dimensionality problem. Therefore, the neural-based sequential learning model is used to achieve superior results for fake news detection. The results are drawn to validate the importance of the linguistic model extracted features and finally combined linguistic feature-driven model is able to achieve the average accuracy of 86% for fake news detection and classification. The sequential neural model results are compared with machine learning based models and LSTM based word embedding based fake news detection model as well. Comparative results show that features based sequential model is able to achieve comparable evaluation performance in discernable less time.

A Grid-Density Based Algorithm by Weighted Spiking Neural P Systems with Anti-Spikes and Astrocytes in Spatial Cluster Analysis

In this paper, we propose a novel clustering approach based on P systems and grid- density strategy. We present grid-density based approach for clustering high dimensional data, which first projects the data patterns on a two-dimensional space to overcome the curse of dimensionality problem. Then, through meshing the plane with grid lines and deleting sparse grids, clusters are found out. In particular, we present weighted spiking neural P systems with anti-spikes and astrocyte (WSNPA2 in short) to implement grid-density based approach in parallel. Each neuron in weighted SN P system contains a spike, which can be expressed by a computable real number. Spikes and anti-spikes are inspired by neurons communicating through excitatory and inhibitory impulses. Astrocytes have excitatory and inhibitory influence on synapses. Experimental results on multiple real-world datasets demonstrate the effectiveness and efficiency of our approach.

Recognizing multivariate geochemical anomalies for mineral exploration by combining deep learning and one-class support vector machine

The recognition of multivariate geochemical anomalies is important for mineral exploration. Big data analytics, which involves the whole data and variables, is an alternative manner to delineate multivariate geochemical anomalies in support of machine learning algorithms due to their strong ability to capture the complex intrinsic and diverse links between geochemical characteristics and mineralization. However, this method faces the issue of data redundancy and calculation complexity, and high-dimensional problems raise great challenges for anomaly detection. This is the curse of dimensionality problem, which hinders the development of a variety of techniques for anomaly detection. In this study, a hybrid model that combines unsupervised deep belief networks (DBNs) and one-class support vector machine (OCSVM) is adopted to address the high-dimensional geochemical anomalies detection problem. In this model, the relevant features first extracted through the DBN are used as the input of the OCSVM. The decision function values of the hybrid method are employed to map the geochemical patterns related to iron mineralization. The comparative results on the performance of the hybrid model and the other three anomaly detection models (deep autoencoder model, OCSVM, and hybrid model with principal component analysis and OCSVM) in terms of the area under curve(AUC) values, suggest that the hybrid method of the DBN and OCSVM can efficiently recognize the geochemical anomalies related to iron mineralization. The DBN can extract the geochemical information, reduce the redundant features, and further enhance the scalability of the OCSVM for processing high-dimensional geochemical data. The extracted geochemical anomalies, which show a close spatial relationship with the Yanshanian intrusions, can provide significant guidance for the next round of mineral exploration.

MDP-Based Task Offloading for Vehicular Edge Computing Under Certain and Uncertain Transition Probabilities

Low latency/delay is one of the most critical requirements for the application of vehicular networks. However, frequent real-time information update caused by vehicles high mobility is liable to aggravate the delay. Meanwhile, the task migration between different vehicular edge computing (VEC) servers results in an amount of delay if the computing cannot be completed before the vehicle moves out of the coverage of the current VEC server. In this paper, the problem is concluded as when and to whom to offload the task for VEC, which is formulated as a finite horizon Markov decision process (MDP) to minimize the delay with respect to the communication, computing, handover and migration. Through characterizing the time-space correlation of vehicles mobility, the curse of dimensionality problem in MDP is resolved. Meanwhile, a general expression of the transition probabilities is derived. On this basis, the specific results of highway, 2-D street and real-data scenarios are provided as well. For practical implementation considerations, the transition probabilities are commonly uncertain primarily due to random driver behavior, inaccurate sample data and complex path environment. Under this uncertain environment,a robust time-aware MDP-based task offloading algorithm (RTMDP) is proposed, which has been proved to perform well even under the high uncertain transition probabilities by simulation results.

R-Ensembler: A greedy rough set based ensemble attribute selection algorithm with kNN imputation for classification of medical data

Background and ObjectiveRetrieving meaningful information from high dimensional dataset is an important and challenging task. Normally, medical dataset suffers from several issues such as curse of dimensionality problem, uncertainty, presence of missing values, non-relevant and redundant attributes, etc. Any machine learning technique applied on such data (without any preprocessing) by and large takes a considerable amount of computational time and may degrade the performance of the model. MethodsIn this article, R-Ensembler, a parameter free greedy ensemble attribute selection method is proposed adopting the concept of rough set theory by using the attribute-class, attribute-significance and attribute-attribute relevance measures to select a subset of attributes which are most relevant, significant and non-redundant from a pool of different attribute subsets in order to predict the presence or absence of different diseases in medical dataset. The main role of the proposed ensembler is to combine multiple subsets of attributes produced by different rough set filters and to produce an optimal subset of attributes for subsequent classification task. A novel n number of set intersection method is also proposed to reduce the biasness during the time of attribute selection process. Before selecting the minimal attribute set from a given data by the proposed R-Ensembler method, the dataset is preprocessed by the k nearest neighbour (kNN) imputation method for missing value treatment. ResultsExperiments are carried out on seven benchmark medical datasets collected from University of California at Irvine (UCI) repository. The performance of the proposed ensemble method is compared with five state-of-the-art attribute selection algorithms, results of which are measured using three benchmark classifiers viz., Naïve Bayes, decision trees and random forest. Experimental results clearly justify the superiority of the proposed R-Ensembler method over other attribute selection algorithms. Results of paired t-test performed on average accuracies produced by different classifiers simulated on the reduced data sets achieved by the proposed and counter part attribute selection methods confirm the statistical significance of the better reduced attribute subsets achieved by the proposed R-Ensembler method compared to others. ConclusionThe proposed ensemble method turned out to be very effective for selecting high relevant, high significant and less redundant attributes from a pool of different subsets of attributes.

Hedging Strategies for Net Interest Income and Economic Values of Equity

The changes in Interest rate risk in the banking book ( IRRBB ) of Basel capital frameworks Pillar 2, where the regulator prescribed a set of stress scenarios and shocks to a standardized interest rate risk measures on changes of Economic Value of Equity (EVE) and Net Income Interest (NII) force banks to model interest rate risks and develop hedging strategies to mitigate the sensitivity to interest rate shocks. In this work we define, model and compute both the NII and the EVE indicators. We study early renegotiation models, based either on an optimal exercise or a statistical model. Finally, we discuss hedging strategies for the NII and the EVE based on both variance and sensitivities minimization. The framework is flexible enough to take into account a wide variety of operational situations, including regulatory ones. To test it, we analyze numerically hedging strategies aiming to reduce the sensitivity of ALM risk measures. Notably, we show that the NII and the EVE are opposite measures: lowering the sensitivity of one of this measures increases the sensitivity of the other. However, anticipated refunds and renegociation effects might temper this conclusion. To reach these goals, we shift from a classical scenario-based definition of the NII and the EVE to a stochastic one, that is quite similar to a Front-Office portfolio analysis. However, there is a drawback in this approach: the targeted applications require not only to compute the EVE and the NII indicators, but also all their conditional futures prices and sensitivities' values. In an operational context, where the number of risk sources can be large, classical numerical methods face the so-called curse of dimensionality problem. To tackle this difficulty, we conduct our numerical experiments using a financial mathematics framework, embedding a partial differential equations (PDE's) solver, that overcomes the curse of dimensionality.

An effective dimension reduction algorithm for clustering Arabic text

Text clustering is a challenging task in natural language processing due to the very high dimensional space produced by this process (i.e. curse of dimensionality problem). Since these texts contain considerable amounts of ambiguities and redundancies, they produce different noise effects. For an efficient and accurate clustering algorithm, we need to extract the main concepts of the text by eliminating the noise and reducing the high dimensionality of the data. This paper compares among three of the famous dimension reduction algorithms for text clustering to show the pros and cons of each one, namely Principal Component Analysis (PCA), Nonnegative Matrix Factorization (NMF) and Singular Value Decomposition (SVD). It presents an effective dimension reduction algorithm for Arabic text clustering using PCA. For that purpose, a series of the experiments has been conducted using two linguistic corpora for both English and Arabic and analyzed the results from a clustering quality point of view. The experiments have shown that PCA improves the quality of the clustering process and that it gives more interpretable results with less time needed for the clustering process for both Arabic and English documents.