International Journal of Computational Intelligence Studies (14 papers in press)
Application of Computational Intelligence Techniques for Internet of
Things: An Extensive Survey
by Shreyas J, Anand Jumnal, Dilip Kumar S M, Venugopal K R
Abstract: The application of computational intelligence (CI) techniques to Internet of Things (IoT) is gaining popularity due to its capability of providing human-like knowledge, such as cognition, recognition, understanding, learning, and others. This paper attempts to provide an exhaustive survey of the available literature on IoT using CI techniques. In addition, detailed categorization has been provided on the basis of different CI tools and their hybridizations used to tackle different problems of IoT. The potential benefits and utility of CI techniques in IoT are highlighted. The possible mapping of CI techniques to the real-world IoT problems is presented. The advantages and disadvantages of CI algorithms over traditional IoT solutions are discussed. A general evaluation of CI algorithms is presented, which will serve as a guide for using CI algorithms for IoT. Finally, some considerations regarding the recent trends and potential research directions are presented. An extensive bibliography is also included.
Keywords: Artificial Immune System; Computational Intelligence; Fuzzy Logi; Genetic Algorithms; Internet of Things; Neural Networks; Swarm Intelligence.
Alzheimer's disease prediction using Regression models and SVM
by Rohini M
Abstract: Alzheimer's disease (AD) and cognitive impairment due to aging are the recently prevailing diseases among aged inhabitants because of an increase in the aging population. Several demographic characters, structural and functional neuroimaging investigations, cardio-vascular studies, neuropsychiatric symptoms, cognitive performances and biomarkers in cerebrospinal fluids are the various predictors for AD. We can consider these input features for the prediction of symptoms whether they belong to AD or normal cognitive impairment for aging. In the proposed study, the hypothesis is derived for supervised learning methods such as multivariate linear regression, logistic regression, and SVM. We perform feature scaling and normalization with features as an initial step for applying the parameters to derive the hypothesis. We analyze performance metrics with the implementation results. The present work is applied to 1000 baseline assessment data from Alzheimers disease Neuro-Imaging Initiative studies (ADNI) that give conversion prediction. The comparison of results in literature studies suggests that the efficiency of the proposed study is highly helpful in differentiating AD pathology from cognitive impairment because of aging.
Keywords: Multivariate linear regression; logistic regression; Support Vector Machine(SVM);Feature scaling; Normalization;ADNI.
Detection of various categories of Fruit and vegetable through various descriptor using machine learning techniques
by Mukesh Tripahthi
Abstract: An accurate and efficient recognition system for fruits and vegetables is one major challenges. To solve this challenges, we have examined various feature descriptor based on colour and texture such as RGB, CMH, CCV, CDH, LBP, CSLBP and SEH. All process of proposed framework consists three phase 1) background subtraction 2) Feature extraction 3) training and classification. In this paper, Otsus thresholding is used for background subtraction. Further all segmented image is used in feature extraction phase. Finally, C4.5 and KNN is used for training and classification. The various performance metric such as CA, Precision, Recall, F-measure, MCC, PRC and FPR are used to evaluate the proposed system for recognition problem. We, also analysis performance accuracy of both classifiers. In that C4.5, KNN classifier produce CA with value 94.63 %, 90.25 % respectively.
Keywords: Detection; fruits; vegetables; descriptor; Performance metric; C4.5; KNN.
Special Issue on: GUCON-2018 Intelligent Models for Emerging Technologies
Empirical Estimation of Various Data Stream Mining Methods
by Ritesh Srivastava, Veena Tayal
Abstract: The online learning is done in order to work in dynamic environments in which concept tend to change with time and the accuracy of classifiers decreases. The current and previous research is done in static environments, but there is a need of real time data streaming due to the potentially larger number of applications available in the scientific and business domains. There are several methods used in learning in the presence of dynamic environments like single classifier methods such as batch and incremental learning approaches, classification methods with explicit drift detection method, windowing techniques and ensemble approaches. This paper, investigates these approaches for determining the best suitable method among them. We utilized light emitting diode (LED) data generator for evaluating the performance of the methods.
Keywords: Concept drifts; online learning; data stream mining; machine learning; classification; drift detection methods.
Special Issue on: BDCA'18 Data Science and Applications
Information Technology performance management by Artificial Intelligence in Microfinance Institutions: An overview
by Kaicer Mohammed
Abstract: This paper presents an overview of the use of new information technology to improve the management of microfinance institutions, experiencing a gap due to the growth of the microfinance sector and the diversity of products and services they offer to the target populations. We will show that artificial intelligence could play a role to ensure reliable management information systems in MFIs.
Keywords: Management of Informatics Technology; Artificial intelligence; Microfinance Institution; Central risk.
Intelligent intrusion detection system using multilayer perceptron optimized by genetic algorithm
by Mehdi Moukhafi, Khalid El Yassini, Bri Seddik
Abstract: This paper presents a neural network-based intrusion detection method for the attacks on a computer network. Neural networks are used to predict unusual activities in the system. In particular, feedforward neural networks with the back propagation training algorithm were employed in this study. we propose a method of intrusion detection based on a combination of GA(Genetic algorithm) and MLP (Multilayer Perceptron) Neural Network to develop a model for intrusion detection system. All tests were realized with the kdd99 data set. The performance of the proposed method of intrusion detection was evaluated on all KDD99 data set, 10% of the KDD99 data set were used for training the GA-MLP model. This system achieves a top accuracy of up to 93.05%.
Keywords: Machine Learning Based Intrusion Detection; Parameters optimization; Genetic algorithm; Multilayer Perceptron Neural Network.
QoE in Video Streaming over Ad-hoc Networks: Comparison and Analysis of AODV and OLSR Routing Protocols
by Hind ZIANI, Nourddine ENNEYA
Abstract: Video Streaming services are easily among of the most consumed services on the internet. Indeed, they are single-handedly accountable for up to 85% of overall internet traffic. And yet, despite the multiple modern infrastructure networks and high-end technologies which remain in constant evolution, network masters still assess Quality levels by its dependent and independent factors. Furthermore, new venues of marketing strategies are constantly witnessing the emergence of ever-novel, ever-revolutionary quality horizons which centralize Human perception. And so, in addition to the Quality of Service (QoS) which is built upon network-oriented metrics, we are now faced with stakes bearing on the Quality of Experience (QoE). In MANET, guaranteeing good quality and performance, be it objective or subjective, is a challenge to be reckoned with. In fact, The extant routing protocols are generally network-oriented and are, as such, chiefly dependent upon objective quality parameters, whence they seldom correlate with the QoE standards as averred by the users perception of the received service. This article purports to analyze and experiment on video transmission, through an Ad-hoc network, based on two emblematic routing protocols -AODV and OLSR- in view of identifying the one most relevant to, and optimal for the subjective quality (QoE) we are focused on.
Keywords: Ad-hoc networks; Video Streaming; Routing protocols; Quality of Experience.
Efficient of Bitmap Join Indexes for optimizing Star Join Queries in Relational Data Warehouses
by Mohammed YAHYAOUI, Souad AMJAD, Lamia BENAMEUR, Ismail JELLOULI
Abstract: Data warehouses are dedicated to analysis and decision-making applications. They are often schematized as star relational models or variants for on-line analysis. Typically, the analysis process is conducted via OLAP (On-Line Analytical processing) type queries. These queries are usually complex, characterized by multiple selections operations, joins, grouping, and aggregations on large tables. Which require a lot of calculation time and thus a very high response time. The performance of these queries depends directly on the use of the secondary memory. Indeed, each input-output on disk requiring up to ten milliseconds. In order to reduce and minimize the cost of executing these queries, the data warehouse administrator must make a good physical design during the physical design and tuning phase by optimizing access to the secondary memory. We focus on bitmap join indexes that share the same resource, that is, the selection attributes extracted from the business intelligence queries. To optimize star join queries.
Keywords: Data Warehouse; OLAP; Indexes; Optimization Query; Star join query; Bitmap join indexes.
Wolf : A framework for digital workplace - Architecture and models -
by Khadija ELAMRANI, Noureddine Chenfour, Mohamed LAHMER, Ghita Daoudi
Abstract: The main purpose of the digital workplace (DW) is to ensure to the organizations different contributors or actors a portal of digital services, which are accessible through a virtual desktop covering all its business services. During our studies, we were able to identify five major problems. First of all, we note a great confusion in the related definitions because most of them are restricted to the teaching sector. Secondly, most existing DWs are summarized as a simple gateway to pre-existing digital tools collection that covers the organizations business domains, without any means of communication between them. Another problem is the lack of a reference architecture. Moreover, we could not identify any logical or physical model to represent the different DWs entities. Lastly, there is a total absence of a standard or even an appropriate vocabulary.rnFaced with these shortfalls, we propose in this paper a set of fundamentals that is composed by a definition encapsulating the different domains, as well as a naming system and a vocabulary that identify both the entities that compose the virtual desktop and their connections and flows. Based on these fundamentals, we also propose our framework WOLF (Digital Workplace based on Open and Light architecture Framework) that generate automatically customized digital workplaces, and is distinguished from other existing DWs solutions by its generic and extensible character. The generated DW encapsulates all of the organizations domains, services, flows and a collaboration system between the different actors. Our proposed frameworks architecture allows us to classify and organize the various entities into a tree representation whilst data nodes are modelled using XML files.
Keywords: Digital workplace; Digital workspace; Collaboration; Digital work environment.
Special Issue on: Intelligent Systems for Cyber Security Current Trends, Applications and New Challenges
Intrusion Detection using Data Mining
by Shubha Puthran, Ketan Shah
Abstract: Intrusion Detection plays very important role in securing Information Servers. Classification and Clustering Data Mining algorithms are very effective to deal with Intrusion Detection. However, classification (supervised) results with false negative detection and Clustering (unsupervised) results with false positive detection. This paper introduces a unique framework consisting of Pre-processing unit, Intrusion detection using quad split(IDTQS), Intrusion Detection using Correlation based quad split (IDTCA) and Intrusion Detection using Clustering (IDTC). In this proposed framework, IDTQS and IDTCA shows accuracy improvement for University of New South Wales (UNSW) dataset is in the range 4%-34% for DoS, Probe, R2L, U2R and Normal classes. IDTC Clustering algorithm performs with 97% accuracy. Training and testing time is improved by 14% for IDTCA in comparison with IDTQS.
Keywords: Quad split; Decision Tree; Correlated Attributes; UNSW dataset.
An Integrated Approach for Multimodal Biometric Recognition System using Pearson Type-II (Beta) Distribution
by Naga Jagadesh Bommagani, A.V.S.N Murty
Abstract: Biometric recognition plays an important role in personnel identity authentication. Usually biometric recognition protocols which involve single source of information are called unimodal systems. Such systems suffer from the problems like noisy sensor data, performance, collectability and non-universality. To have an accurate recognition it is needed to develop a system with multimodal biometrics. Hence, in this paper a new approach is proposed with the combination of multiple biometric traits such as face, fingerprint and palm vein. Region of Interest (ROI) is used to consider the valuable information from the images. The 2D Discrete Cosine Transform is used for extracting the feature vector from face, fingerprint and palm vein and fusion at feature extraction level. Here the feature vector is modelled with Pearson Type-II distribution and the model parameters are estimated using the EM algorithm. The initialization of model parameters is done through moment method of estimators and K-means algorithm. The performance of the proposed algorithm is carried by experimentation with CASIA biometric database. Through experimentation the proposed model performs more effectively than the algorithm with Gaussian mixture model.
Keywords: Multimodal biometric recognition; Discrete Cosine Transform; EM algorithm; Pearson mixture model.
IbPaKdE: Identity-Based Pairing free Authenticated Key and Data Exchange Protocol for Wireless Sensor Networks
by Lakshmana Rao Kalabarige
Abstract: The security vulnerabilities in key distribution approaches of WSN
reveals important credentials. The secure distribution of keys without having
permanent storage of important credentials in the permanent memory part
(ROM) of a sensor node is a challenging task. Further, the design of a key
distribution approach with less computational complexity, energy efficient, low
communication overhead, and low memory overhead are some challenging
tasks for a resource constrained sensor nodes. This piece of work addresses
all these challenges by combining Identity-Based Cryptography(IBC) with
Symmetric Key Cryptography(SKC). The proposed Identity-Based Pairing free
Authenticated Key and Data Exchange Protocol(IbPaKdE) avails advantages
of both IBC and SKC to address the above challenges. This approach does
not require prior communication for the establishment of secret keys and it
supports pairing free key distribution. The proposed IbPaKdE uses IBC strategy
for secure exchange of keys and SKC to provide security to the data to be
transmitted. The on demand establishment of keys eliminates the permanent
storage of important credentials in the sensor nodes. Simulation results of the
proposed approach is compared with hashed identity based secure key and data
exchange(HISKDE). The results show that the IbPaKdE incur better results than
HISKDE in terms of energy efficiency, reduces memory, communication and
computational, overheads of a sensor node.
DCT statistics and pixel correlation based blind image steganalysis for identification of covert communication
by Madhavi B. Desai, S.V. Patel, Vipul H. Mistry
Abstract: In the last decade, the interconnection of systems through networks, access to information, different computer technologies and the combination of all these aspects have increased the use of image steganography techniques for illegal acts. Furtherance of image steganography techniques exploited to send secret information on social network builds the requisite of blind image steganalysis. Blind image steganalysis is one in which no prior information is available about the data hiding method used to embed the message. Existing image steganalysis methods are either domain specific or the one which requires a very high dimensional feature set. Considering the types of image steganography methods, embedding rates, image types and feature dimensionality there is an utmost need of a low dimensional blind image steganalysis method. This paper proposes a blind steganalysis method with a 32-D feature set comprising of DCT Statistics and Pixel Correlation (DSPC) algorithm with the aim to the reduced time complexity of feature extraction as well as the complexity of classifier. The experimental results evidence that the proposed feature set gives better results against state-of-art high dimensional image steganalysis methods. The performance of the proposed algorithm is evaluated using experiments with varying embedding message size, message types and image formats using Ensemble classifier. The algorithm is implemented in Matlab and all the experiments are performed on standard image datasets i.e. BSDS300, CorelDraw.
Keywords: Blind Image Steganalysis; Binary Similarity Measures; DCT Transform; Ensemble Classifier; Feature Extraction; Statistical Features; and SVM.
Adaptive QoS Constraint-based Service Differentiated Routing In Wireless Sensor Network
by Yogita Patil
Abstract: Achieving the best quality of service (QoS) as per user requirement is one of the important challenges. The time-critical applications in Wireless Sensor Networks (WSN) demand energy efficient transmission of data with limited resource availability. To resolve these issues AQSDR has been proposed. The proposed protocol support packet differentiation and selection of sensor node based on energy, delay, and congestion for path establishment to transfer normal data packets providing energy conservation. The multipath is chosen for transmission of emergency packets satisfying delay requirements of non-delay tolerant applications. AQSDR support adaptive path selection, according to application requirement. The proposed technique in this work outperforms when compared with the existing protocol in terms of minimized energy consumption, delay, control overhead, packet drop ratio, and high throughput.
Keywords: Clustering; Congestion Index; Delay; Energy Efficiency; Multipath; Service differentiation; WSN.