Iraqi Journal for Computers and Informatics

Over time, with the growth of credit cards and financial data,credit models are needed to support banks in making financial decisions. Hence, developingan efficient fraud detection systemis essential to avoid fraud in Internet transactions,which increased with the growth of technology. Deep learning techniques are superior to other machine learning techniques in predicting the behavior of credit cardcustomers basedon the probabilitythat they will miss a payment. The bidirectional long-short term memory (BiLSTM) model is proposed to train the Taiwanese non-transactional dataset for bank credit cards to decrease the losses of banks. The BiLSTM reached an accuracy of 98% in fraud credit detection compared with other machine learning techniques.

Alcoholism is a severe, disorder that affects; the functionality of neurons in the central nervous system
and leads to the loss of. health and wealth. The suggested technique applies statistical and fractal
dimension (FD) features to classify alcoholic and normal subjects using eight channels under an SFbased
machine learning architecture. Electroencephalogram (EEG) signals are placed in a framework
and separated into different EEG bands using an orthogonal wavelet filter. The following three
classification approaches are used to classify the alcoholic and normal patterns of EEG data: leastsquare
support vector machine, vector machine (SVM), and Naïve Bayesian. Results showed that the
best classification method was SVM with a sensitivity of 0.9267%, an accuracy of 0.9892%, and a
specificity of 0.9916%.

Algorithms for string matching are considered one of the most extensively researched topics in the field of computer science due to theirsubstantial role in various applications, such as information retrieval, editing, security, firewalls, and biological applications. String matching involves examining the optimal alignment by comparing the characters in the pattern and the text. Over the past two decades, it has gained considerable attention due to technological advancements. The need to address string-matching problems has also emerged because of its wide-ranging applications. This study presents the E-ARFO hybrid string-matching algorithm, which combines the best features of two original algorithms, namely, E-AbdulRazzaq and fast online hybrid matching. Compared with other algorithms, the proposed method demonstrates outstanding performance in terms of the number of attempts and charactercomparisons conducted across multiple databases, including DNA and protein sequences. Results indicate that irrespective of the number of attempts or character comparisons made, E-ARFO consistently ranks first for short and lengthy patterns in most databases. Results also reveal reduced runtimes and competitive character comparisons. Moreover, results underscore the potential effect of E_ARFO on computational biology, offering a new paradigm for precision and efficiency in string matching.

Researchers' interests and academic journals are crucial for advancing scientific inquiry. Journals serve as platforms for sharing and validating discoveries, fostering a symbiotic relationship that advances our collective understanding and pushes the boundaries of human knowledge. Journals, which encompass natural edge research and establish benchmarks for academic rigor. In this paper, an analysis, using text mining, of the publications of Iraqi researchers in scientific journals is used to extract the researcher's interest. In more detail, this paper utilizes the following technologies: pre-processing (tokenization, POS (“Part Of Speech”), normalization, case folding, lemmatization) –filtering (stop word elimination) -feature Extraction (TF-IDF), aswell as classification using deep neural network classifier (DNNC), to address the problem of identifying the researcher's interests through texts (title &abstract) analysis. The Iraqi researchers’ data in the field of computer science from the years 2010-2022. As obtained from the Scopus repository, a total of 1170 papers were collected via API-key and scrubber depending on the keyword of computer science and the year. Furthermore, these papers were manually classified based on the hierarchical classification of the ACM journal. Finally, the best results obtained from a classification using DNN and TF-IDF as classifying terms achieved a precision of 90%, Recall of 90%, f1-score of 90%, and accuracy of 90%.

Thisinvestigationaimstoexplorepreviousresearchontheimplementationoffeatureselectioninintrusiondetection.Featureselectionhasdemonstrateditsabilitytoenhanceorsustaincomparableclassificationaccuracylevelsforintrusiondetectionsystems,whilesimultaneouslyimprovingclassificationefficiency.Theevaluationincludesanassessmentoffilter-based,wrapper-based,andhybridfeatureselectiontechniques.GiventhatBigDatachallengescanaffectintrusiondetection,featureselection’sclassificationefficiencycanaidinloweringcomputingrequirements.OlderKDDintrusiondetectiondatasetshavereceivedconsiderableattentioninpreviousfeatureselectionresearch.Consequently,researchersneedmorehigh-qualitydatasetsthatareavailabletothegeneralpublic.

With the rapid proliferation of Internet of Things devices in smart homes, the necessity of robust security measures has become clearer than ever. This study aims to apply data analytics to enhance security monitoring within smart home environments. By leveraging the wealth of data generated by IoT devices, the study also aims to create an intelligent system that is capable of proactively identifying and addressing potential security threats through Android mobile devices.Therefore, privacy concerns were addressed through encryption and anonymization methods to protect sensitive information. The study evaluates the effectiveness of the developed security monitoring system through simulated and realistic scenarios, thereby highlighting its ability to detect and mitigate a wide range of security threats. Thus, the research contributes to the development of smart home security by providing a smart, data-driven approach to monitoring security incidents. In the evolving landscape of smart homes, the proposed framework forms a cornerstone for ensuring the safety and privacy of residents within this interconnected ecosystem.

Crime is undeniably rising, thus negatively affecting countries’ economies. Despite several efforts to study crime prediction to reduce crime rates, few studies take the timeline factor into account when extracting crime-related tweets to predict crime. Aimingto predict Arabic crime tweets on Twitter/X, this study predicts crimes after analyzing social sentiment—that is, whether a tweet raises positive, negative, or neutral feelings—and filters the tweets based on crime behavior through an intelligent dictionary built through a genetic algorithm. The study uses a variety of machine learning (ML) models—random forest, logistic regression, and decision trees—which are assessed according to their accuracy, precision, recall, and F1 scores to guarantee robustness and dependability in crime prediction. The accuracy after filtering crimes based on an intelligent dictionary are 97% for decision tree, 97% for random forest, and 94.43% for logistic regression. This research will provide insight into potential crime attitudes and public opinion toward safety and law enforcement

Numerous sensor nodes make up a wireless sensor network. Sensor nodes work together to send information about the environment to the base station. Nodes are susceptible to failure during these operations due to power failure, hardware or software failure, etc.. Consequently, delivering dependable packages necessitates achieving two important objectives: energy efficiency and minimized error rate. A novel approach is proposed to distribute the workload among each sensor node to accomplish these objectives. When estimating the scope of communication between nodes, uncertainties are dealt with using the fuzzy logic method. During some weather hazards, the features and cables are particularly challenging to monitor with simple devices in the WAN area. Remote sensing units are used to check and display the feature conditions to resolve this issue. Using machine learning algorithms, the collected data will be analyzed and processed to create a prediction horizon that will be useful for environmental studies and early warning systems. The MatLab2020b simulation program, which has useful tools and learning applications, willbe used to carry out this task. The FIS algorithm obtained a cost distance of just 629 m out of the total WSN sensor lengths of 2600 m, which is the best cost among all implemented routing protocol techniques in this study.

This study investigates how decentralization and transparency offered by blockchain technology could revolutionize traditional finance. Even with the rise of well-known cryptocurrencies such as Bitcoin and Ethereum, a general understanding of blockchain’s influence on the financial industry is still lacking. We identified five major application cases—transparent credit scoring, effective consumer identification, expedited insurance settlements, improved cybersecurity, and the emergence of decentralized finance—where blockchain technology is well positioned to tackle persistent issues. We show how blockchain technology may address problems such as opaque credit scoring, poor customer identity, convoluted insurance settlement procedures, and susceptibility to cyberattacks by thoroughly examining various use cases. According to our research, a greater number of traditional financial institutions need to embrace and integrate blockchain innovations into their functions to promote inclusivity, transparency, and decentralization.

Thisstudydescribesalightweightdeeplearningmodeltrainedonaself-madeimagedatasettakeninsidefarmsandopenareasoftheHolyShrineofAl-HussainiyaintheCityofKarbala,Iraq.ThisdatasetincludesfireandsmokeimagestakenusingaSamsungA52Scameraindifferentweatherconditions.Theoverallgoalistocreateafiredetectionsystemmodelthatcansuccessfullyreplacetheexistingphysicalsensor-basedfiredetectorsandlessentheissuesthatcomewithsuchfiredetectors,includingfalseanddelayedtriggering.Anothergoalistocontrolfiresonfarmsoropenareasandpreventcropdamageasmuchaspossible.Previousstudieswerereviewed.Moreover,thearchitectureoftheYouOnlyLookOnceversion8(YOLOv8)modelwasbrieflyexplained,andtheresultsitachievedwerecomparedwiththoseachievedbypreviousversions.Then,theproposedsystemwastrainedandevaluatedwiththeYOLOv8largemodel.ResultsshowedthattheproposedsystemoutperformedtherestofthecurrentsystemsinmAP,whichreached98.5%ز

In the present era of smart grids, accurate prediction of energy uses is becoming increasingly essential to guarantee optimal energy efficiency. This study contributes to the field by utilizing advanced machine learning techniques to perform predictions of energy consumption using the data from Internet of Things (IoT) devices. Specifically, the approach utilizes regression neural network (RNN) structures, such as long short-term memory (LSTM) and gated recurrent units (GRUs). The data from IoT sensors are more extensive and detailed than those of conventional smart meters, allowing for the development of more complex models of energy use patterns. This study utilizes Adam-optimized LSTM, RNN, and GRU models, along with stochastic gradient descent, to evaluate their performance in addressing the complexity of time-series data in energy forecasting on different network configurations. Result of the analysis indicates that LSTM models, which are run with the Adam optimizer, are more accurate in terms of predictions compared with the other models. This conclusion is supported the test results of these models that are within the lowest root mean square error and mean absolute error scores. All the models under the analysis exhibit signs of overfitting based on the performance indicators for the training and the testing data. This notion implies that the regularization should be utilized to ensure the improved generalizability of the models. Thesefindings show that deep learning can have a lasting influence in improving energy consumption management systems to meet the sustainability and energy efficiency requirements. These observations are beneficial for the gradual improvements of smart grids.

Numerous people have died as a result of the coronavirus outbreak in 2019 (COVID-19), which also affected millions of others around worldwide. The infection spreads quickly. Therefore, technology that enables quick virus detection will offer healthcare professionals much-needed assistance. This study aims to identify COVID-19 disease from X-ray images of healthy and infected people with pneumonia by using a modified VGG16 model. The proposed model achieved better results than previous studies presented with an accuracy of 99.13%, a recall of 99% and a precision of 98.70%.

ThisarticlediscussestheevidentiaryrequirementsfordemonstratingscientificnegligenceunderGhana’sCivilProcedureRules2004(C.I.47)inthecontextofemergingartificialintelligence(AI)diagnosticandtreatmentstructures.Method:Legalanalysisexaminesgapsinsatisfyingburdenofproofandstandardsofevidence,obstaclesthatrestrictevidencecollectiononAIdevicedeficiencies,andsuggestionsforadaptinglegalresponsibilitypoliciestoAI’stechnicalopacity.Findings:Thepresentinabilitytointerrogatealgorithms,limitedaccesstoproprietarytrainingdataandmethods,lackofdiagnosedstandardsofcareforsoftware-baseddecision-makers,andshortageofqualifiedprofessionalwitnessesposemassiveevidentiarychallengesforplaintiffsseekingtoconfirmAInegligence.Conclusions/Recommendations:Standardsstrengtheningalgorithmictransparency,auditability,andexplainabilitycouldeaseevidentiaryburdensforaffectedpatients.StrictliabilityschemesandIPprotectionsbalancingpublicsafetyandinnovationaimsneedtobeconsideredmovingforward.ScientificContributions:Thisworkadaptstraditionalmedicalliabilitysystemstotoday’srealitiesofincreasingrelianceonAIinhealthcareandproposesseveralimprovements.

Withtheadvancementofcommunicationtechnology,manyinnovativeapplicationshavedevelopedinagricultureasaresultoftheintegrationoftheInternetofThings(IoT)withunmannedaerialvehicles(UAVs),leadingtothemodernizationofagriculture.Thisstudyseekstoproposeaneffectiveandlow-costplatformforcomprehensivemonitoringofenvironmentalparametersusingIoTanddrones.ThepreparationofthispaperwasbasedonaplatformthatwastestedinarealisticenvironmentonafarmnearthecityofAl-MedianinTunisia,wheretheplatformwasbuilttosuittherealisticenvironmentofafarminBaghdad,throughtheuseofsensorsaboveandbelowtheground,whichmeetstheexperimentalworkandstandardsforautomatedandreal-timemonitoring.Forenvironmentalstandards,theunifiedtheoryofacceptanceanduseoftechnologymodelwasused,whichisamodelbasedonfourbasicelements:1)expectedperformance,2)expectedeffort,3)socialimpact,and4)facilitatingconditionsforobtainingresults.TheuniqueintegrationofIoTsensorswithdroneshasshownimpressiveexperimentalresults,indicatingthepossibilityofperformingbothautomaticandmanualactionsbyhumans.Thesesmartmovescontributesignificantlytopromotingprecisionagriculture,leadingtoasignificantincreaseinagriculturalproductionandconservationofnaturalresources.

The objective of this paper is to propose an offline signature verification system(OSVS) designed for writer-independent applications. The system shall differentiate between original and forged signatures. The system encompasses key stages such as pre-processing, feature extraction, and model training and testing, employing the SupportVector Machine algorithm for classification. One challenge in creating a good OSVS is to find proper signature features to be used in the training/classification phases. In this research, global and local features are utilized. This includes signature area, mean, standard deviation, perimeter, number of connected components, number of vertical and horizontal edges, number of end points, number of branch points, and number of lines. The contributions of this paper are on several aspects of the offline signature verification process. Investigation in this study include data pre-processing techniques (normalization vs. standardization), kernel selection (Poly vs. RBF), dataset distribution for training and testing (80%-20% vs. 5-fold), and variations of the C and gamma parameters (C=1, 10, 100 and gamma=1, 10, 100). Improving the recognition rate involves removing of features with little or bad effect on the recognition rate. An algorithm for model-agnostic feature importance is executed, revealing that the most crucial features in the classification process are mean, standard deviation, perimeter, number of connected components, and number of end points. Signatures are classified as either original or forgery, and the model's performance is assessed on the CEDAR dataset. Experimental results shows a 95.65% accuracy of the proposed system when utilizing standardization with the RBF kernel.

Thisresearchconductsacomprehensiveinvestigationoftheefficacyofvariousmachinelearningalgorithmsforfiredetection.Thealgorithmsthatwereexaminedincludelogisticregression,decisiontree,randomforest,supportvectorclassifier,gradientboosting,K-nearestneighbors,GaussiannaiveBayes,multilayerperceptronclassifier,andXGBoostclassifier.Throughin-depthexperiments,thisstudyrigorouslyassessestheperformanceofthesealgorithmsinidentifyingandpredictingfiresbasedonpertinentinputfeatures.Amongthealgorithmsthatwereinvestigated,logisticregressionisthebestperformer,withahighaccuracyrateof99%.Thefindingsfromthisresearchoffervaluableinsightsforoptimizingfiredetectionsystems,providinganuancedunderstandingofthepracticalapplicabilityofmachinelearningtechniquesinreal-timefiremonitoringscenarios.Theprimaryobjectivesofthisstudyaretoelucidatespecificchallengesinfiredetection,evaluatetheperformanceofvariousmachinelearningalgorithms,andcontributetothefoundationalknowledgethatisessentialforenhancingfiremanagementstrategies.Theresearchaddressesthelimitedprecisionofexistingfiredetectionsystemsandaimstorectifythisissuethroughasystematicexplorationofadvancedmachinelearningapproaches.Theoverarchinggoalistobolsterthefoundationsoffiremanagement,facilitatingthedevelopmentofproactivemeasuresandpromptresponsestomitigatetheprofoundimpactofwildfires.Bypresentingadetailedexaminationofthestrengthsandweaknessesofvariousmachinelearningalgorithms,thisresearchstrivestofosterarobustandeffectiveapproachtofiredetection,therebyadvancingthefieldand.

A reliable image copy–move forgery detection approach adaptable to different scenarios of tampering with color images is crucial for many applications. Different methods and solutions have been effectively proposed, but they are still subject to false positive/negative detections and cannot handle the variety of copy–move forgeries. In this paper, a machine learning model that combines ResNet 50 and U-net architectures for automatic image forgery detection in color image(s) is presented. The proposed system is inspired by the ResNet 50 architecture as an encoder and the U-Net architecture as a decoder. The encoder function implies applying convolution and normalizing for feature extraction. Conversely, the decoder functions is locating the spatial features. The decoder in the U-Net network comprises multiple decoder blocks, which are connected to corresponding encoder blocks by employing concatenate layers. A binary mask is then produced to represent the tampered regions in the image. Quantitative experimental results on two standard public datasets and a comparison with state-of-the-art methods demonstrate the effectiveness and robustness of the proposed model.

Traditional authentication methods, such as the use of passwords and fingerprints, are susceptible to the risks of theft, loss, and forgery. However, an innovative and secure alternative exists in the form of electroencephalogram (EEG)-based authentication systems, which operate by measuring distinctive brainwave patterns. This particular review undertakes a comprehensive analysis of the current state of EEG-based authentication, delving into its advantages, challenges, and potential future directions. In doing so, we examine the underlying principles governing the acquisition and processing of EEG signals, explore the various techniques employed for feature extraction and classification, and evaluate the performance of existing systems. Moreover, we emphasize the significant advantages offered by EEG-based authentication, including its exceptional accuracy, capacity forliveness detection, and robust resistance to spoofing attempts. Nevertheless, we must also acknowledge and address the various obstacles that must be overcome to facilitate wider adoption of this authentication method, encompassing concerns relating to hardware affordability, user acceptance, and data privacy. Finally, we outline a series of promising research avenues that can potentially address these challenges and unlock the complete potential of EEG-based authentication, thereby enabling secure and convenient access control across a diverse range of domains.