International Journal of Intelligent Engineering Informatics (10 papers in press)
Real-Time Image Encryption and Decryption Methods based on the KarhunenLoeve Transform
by Hasan Rashaideh, Ahmad Shaheen, Nijad Al-Najdawi
Abstract: The need for a reliable, fast, and accurate encryption algorithm that ensures an identical decrypted image is required in many organizations and industries where marginal concessions are not acceptable. The choice of encrypting images in the frequency domain is more suitable as the image components will be de-correlated, which provides a suitable platform to classify those values according to their significance. In this work and based on the Karhunen-Loeve transform and the Advanced Encryption Standard algorithm a symmetric lossless encryption and decryption algorithms are proposed, where various frequencies are processed in order to provide reliable and secure form. The encryption algorithm involves diminishing the horizontal, vertical and diagonal frequencies that represents the visual details of the given image, this is done using a series of operations designed specifically for this purpose. Simultaneously the frequency values are used to build a decryption key matrix that is to be further encrypted using the Advanced Encryption Standard algorithm. On the other hand, the decryption algorithm overturns the encryption steps returning the image to its original shape with no changes on its corresponding values. The proposed encryption scheme is applicable for applications that require protected delivery of high-quality data. In order to evaluate the efficiency of the proposed algorithms, a comprehensive analysis and evaluation has been conducted based on a set of standard benchmark test images. In terms of quality, the proposed decryption algorithm outperforms the rest of the algorithms in this domain by providing lossless decrypted images that are identical to the originals. Moreover, well-known statistical tests and security measures have been performed in order to validate the security of the proposed encryption algorithm, the results outperform the state-of-the-art algorithms in this domain.
Keywords: Cryptography; Lossless Symmetric Encryption and Decryption.
Real Time Fuzzy Logic Controlled Fire Detection System for Home Applications
by Mehmet Cunkas, Vacip DENIZ
Abstract: In this study, a low-cost alternative method for hardware implementation in fire detection has been developed for home or office applications by using fuzzy logic. The system consists of smoke, flame, heat and humidity sensors which are used to detect the fire. When a possible fire is detected, the data from the sensors are processed with fuzzy logic and the result is sent as a message to the mobile phone through the GSM module indicating the fire location in minimum time period. In addition, the sensor values can be monitored by the MQTT Dash application on the mobile phone. A fire test cabinet has been designed to test the developed method. Real time tests were done in the fire test cabinet and it was observed that the fire status was accurately detected in a short time. This study, which is important for researchers and practitioners, offers a different approach to fire detection, especially a new framework for home and office applications, and also helps fire-fighters to respond quickly to fire.
Keywords: Fire detection; fuzzy logic; GSM communication; home applications; sensor monitoring.
Forecasting Time Series Data Using Moving-Window Swarm Intelligence-Optimized Machine Learning Regression
by Ngoc-Tri Ngo, Thi Thu Ha Truong
Abstract: An accurate forecast of future time series data can support decision-makers to obtain economic benefits. This study proposes a hybrid time series forecast model namely a moving-window firefly algorithm (FA)-based least squares support vector regression (MFA-LSSVR). The LSSVR captures patterns of historical data and predicts future values of time series data while the FA is used to optimize the LSSVR`s parameters to improve the predictive accuracy. The proposed model was trained and tested using two actual datasets of the daily energy demand data and the stock price data. Experimental results show that the proposed MFA-LSSVR model is effective in forecasting time series data and the comparison results revealed that the proposed model outperforms other models, i.e., the LSSVR and the ARIMA (autoregressive integrated moving average) in predicting energy demand and stock price. This studys findings, thus, provide decision-makers a potential approach in early forecasting future patterns of time series data.
Keywords: Machine learning regression; moving-window concept; swarm intelligence; time series forecast.
Detection of Threatening User Accounts on Twitter Social Media Database
by Asha Kumari, Balkishan
Abstract: In this technical era, online social media platforms such as Twitter, Facebook, WhatsApp, WeChat, QZone, etc. are instrumental to provide global human connectivity. These social platforms have provided access to the user to the extent that they can fearlessly post and generate a huge amount of data. Although this data is useful to generate useful information, the database contains many of the malicious and threatening users that post the suspicious and fake content on the social media for the personal or organizational advantage. This demands to generate a system that can detect suspicious content and their respective user accounts. In this paper, an Ant Colony Optimization based system for Threatening Account Detection (ACOTAD) is proposed. Ant Colony Optimization helps to evaluate the connections among different twitter users. The connections among the different Twitter users are determined by the pheromone quality among the edges of the path traveled by individual artificial ants. Better the quality of pheromone indicates the strong connection of one user with another. This research work considers the experimentation on Twitter based Social Honeypot database. The feature set related to Twitter users, their accounts, their publically posted tweets, and their connections are considered for evaluation. The results of the proposed system are determined in terms of precision, recall, f-measure, true positive rate, and false positive rate. The calculated results in terms of evaluation metrics indicate the superiority of proposed concept in comparison with state of art techniques.
Keywords: Online Social Media; Twitter; Suspicious Activity; Threatening Users; Ant Colony Optimization; Swarm Intelligence; Twitter Microblogs.
ELECTRONIC SERVICE QUALITY MEASUREMENT (eSQM); DEVELOPMENT OF A SURVEY INSTRUMENT TO MEASURE THE QUALITY OF E-SERVICE
by Hamed Taherdoost
Abstract: Service quality is increasingly recognized as an important aspect of online business activities and is considered as a key determinant for successful electronic service (e-service) as well. Even though, e-service quality has increasingly attracted the attention of researchers, the existing research in this area identified quality dimensions of e-service without deeper attention for development of a new measurement instrument. The absence of a valid and reliable instrument to measure e-service quality in online environment has, however, bedeviled the endeavors of both scholars and practitioners. Therefore, there is a clear need to address the gaps by a theoretical survey tool which integrates all aspects of e-service quality to be deployed in information systems research. In order to respond to the need, the purpose of this article is to examine the theoretical foundation of e-service quality and to develop a survey instrument to measure the quality of e-service. Given the exploratory nature of this research, all characteristics of e-service quality have been extracted from the previous studies. Then, the exploratory factor analysis was used to cluster the factors effectively, thereby, further analysis including content, discriminate, and convergent validity of the proposed survey instrument were tested. The contribution of this research relates to the fact that the proposed theoretical survey instrument integrates in a holistic way various relevant factors affecting e-service quality into a single template and develop E-service Quality Measurement (eSQM) tool.
Keywords: E-service Quality; Service Quality; Survey Instrument; Questionnaire; Exploratory Factor Analysis; E-service Quality Measurement (eSQM).
Optimal self-tuning decoupled sliding mode control for a class of nonlinear systems
by Mohammad Javad Mahmoodabadi, Seyed Mehdi Mortazavi Yazdi, Mohsen Talebipour
Abstract: Decoupled Sliding Mode Control (DSMC) is well-known as a simple way to achieve asymptotic stability of the nonlinear under-actuated systems. This method has many advanced features such as good performance and robustness against parameter variations. However, because of changing the system states, the controller with the constant parameters would not be optimum in any state of the system, and designing DSMC requires an adaptation for the controller parameters. Hence, in this paper, a Robust Self-tuning Decoupled Sliding Mode Controller (RSDSMC) is presented and optimized with five conflicting objective functions and twelve design variables using a multi-objective genetic algorithm. The proposed controller is applied to a highly nonlinear inverted pendulum and cart system via the computer simulation. The final results depict the appropriate performance of this new controller and demonstrate its superiority in comparison with those reported in literature.
Keywords: Robust Self-tuning Controller; Decoupled Sliding Mode Controller; Genetic Algorithm; Inverted Pendulum and Cart System.
Performance Analysis of SEP, I-SEP, PSO and WCA based Clustering Protocols in WSN
by Ankit Gambhir, Ashish Payal, Rajeev Arya
Abstract: Wireless sensor networks (WSNs) are significantly resource restrained by their delimited power supply. Hierarchical routing based on clustering is an effectual methodology; however, preferring an energy-sensible cluster by optimally selection of cluster head (CH) is also very perplexing. In the present paper, performance analysis of the hierarchical routing protocols based on clustering such as stable election protocol (SEP), improved stable election protocol (I-SEP), conventional LEACH (Low-energy adaptive clustering hierarchy) and optimized versions of LEACH based on nature motivated optimization algorithm such as particle swarm optimization (PSO) (derived from swarm intelligence) and water cycle algorithm (WCA) (derived from observing how rivers and streams flow to the sea) are presented. Numerous parameters have been well-thought-out as the base of the study to present useful results in comparison of clustering protocols in WSN that lead to noteworthy findings in the form of results. The study is examined over a variety of performance metrics such as alive nodes per round, packet to base station (BS) per round and dead nodes per round. Based on the thorough consideration and observations, a qualitative conclusion is also drawn at the end.
Keywords: improved stable election protocol; LEACH; particle swarm optimization; stable election protocol; water cycle algorithm; wireless sensor network.
A survey on Database Intrusion Detection: Approaches, Challenges and Application
by Rajni Jindal, Indu Singh
Abstract: Databases store vital information of an organization and are therefore integral for its efficient working. This necessitates the establishment of Database Intrusion Detection Systems (DIDSs) which can detect and prevent unauthorized user access to the critical information stored in database.A lot of work has been done in the field of DIDSs which has grown at a very rapid pace. A large number of publications emerging every year to further improve upon the existing state of the art solutions. This paper investigates research on major approaches proposed in the field of database intrusion detection and analyzes the drawbacks of the proposed methods in order to drive future research towards more efficient and effective DIDSs. A systematic survey is conducted in order to classify various approaches for detecting intrusion in databases. The work identifies open research questions and challenges, by methodically comparing existing strategies to combat malicious transactions in a database system, and also provides an insight to the applications of DIDSs.
Keywords: Database Security; Database Intrusion Detection; Insider Attack; Role Based Access Control; Data Dependency Mining;.
Attention-Based Word-Level Contextual Feature Extraction and Cross-Modality Fusion for Sentiment Analysis and Emotion Classification
by Mahesh Huddar, Sanjeev Sannakki, Vijay Rajpurohit
Abstract: Multimodal affective computing has become a popular research area, due to the availability of a large amount of multimodal content. Feature alignment between the modalities and multimodal fusion are the most important issues in multimodal affective computing. To address these issues, first the proposed model extracts the features at word-level and forced alignment is used to understand the time-dependent interaction among the modalities. The contextual information among the words of an utterance is extracted using bidirectional LSTM. Further bidirectional LSTM is used to extract the contextual information between the nearby utterances before fusion. Weighted pooling based attention model is used to select the important features within the modalities and importance of each modality before fusion. Initially, two-two modalities are fused and then all modalities using cross-modality fusion technique. The performance of the proposed model was tested on two standard published datasets such as IEMOCAP and CMU-MOSI for sentiment analysis and emotion classification respectively. By incorporating the word-level features, feature alignment, and cross-modality fusion, the proposed architecture outperforms the baselines in terms of classification accuracy.
Keywords: Affective Computing; Attention Model; Contextual Fusion; Cross-Modality Fusion; Feature Alignment.
A Knowledge-based Diagnosis Algorithm for Broken Rotor Bar Fault Classification Using FFT, Principle Component Analysis and Support Vector Machines
by Hayri Arabaci, Mohamed Ali Mohamed
Abstract: Despite their ruggedness and reliability, induction motors experience faults due to stresses and manufacturing errors. Early detection of these faults is important in preventing further damages and minimizing down-time. In this study, a machine learning algorithm is proposed for detection and classification of broken rotor bar faults according to their severity. Removal of high frequency components then amplification was performed on the measured single-phase current. Features were then extracted using FFT and Principal Component Analysis (PCA). Support Vector machines (SVM) was used for classification. 2 classification schemes were analysed; one classifying in 1 step and another in 2 steps. Experiments were performed to evaluate the algorithms by analysing their recognition rates. 6 different SVM kernels were studied. Recognition rates as high as 97.9% were achieved. False negative rate as low as 0% was also realized. Furthermore, it was found out that using more principle components does not yield significant improvements.
Keywords: squirrel cage induction motor; IM; support vector machines; SVM; principal component analysis; PCA; BRB; fault diagnosis; machine learning.