Al-Iraqia Journal for Scientific Engineering Research

This study focuses on evaluating the data routing protocol in the Internet of Things (IoT). The evaluation is conducted using the Whale Optimization Algorithm (WOA), enhanced by the advanced Blowfish encryption algorithm. In this study, Cipher Block Chaining (CBC) is employed to improve the typical Blowfish technique. Furthermore, several key factors, including the rate of energy consumption, trust, distance, and delay, are considered to ensure the effectiveness of data routing. To demonstrate the feasibility of the proposed approach, four other schemes are utilized in a comparative study. These schemes are particle swarm optimization (PSO), secure routing protocol ( SARP),Genetic algorithm (GA), QoS-based Cross Multi-sink Routing protocol (QCM2R). The results clearly demonstrate the feasibility of the proposed approach when compared to the aforementioned schemes, with the lowest energy consumption, distance, and delay recorded as 0.44325 Joule, 0.43257 meters, and 1.35 x 10-9 seconds, respectively. Additionally, the trust level is the highest at 0.71255, and the throughput reaches its peak at 2.30 Megabytes/ms when the data size is 7 KB. In terms of security dimension, the proposed improved Blowfish encryption technique is compared with Advanced Encryption Standard (AES), Rivest–Shamir–Adleman (RSA) and typical Blow fish algorithm (BFA) encryption techniques to assess its security robustness against Chosen-Ciphertext Attack (CCA) and Chosen-Plaintext Attack (CPA) with varying the size of data bits from 5KB, 7 KB and 10KB. The obtained results show that the proposed improved blow fish approach outperforms three aforementioned schemes with the smallest similarity between unique text and hacked text under CCA and CPA. Furthermore, the least time of decryption occurs with the proposed approach compared to other three schemes.

This study investigates the performance of an optimal routing protocol in Wireless Sensor Networks (WSNs) based on the Grey Wolf Optimizer (GWO), which is considered an outstanding Metaheuristic optimization technique. Unlike numerous previous studies, this research develops an advanced approach by optimizing the cluster head nodes and sink node mobility, which are essential elements for achieving an optimal routing protocol in WSNs. Several crucial factors, including the rate of energy consumption, and others are considered to ensure trustworthy routing. To demonstrate the feasibility of the proposed approach, three other methods are utilized in a comparative study. Various evaluation measures are exploited, including the number of alive nodes, mean throughput, probability of avoiding unsuitable cluster heads, energy consumption, and computation cost. The simulation results clearly reveal the superiority of the GWO compared to the aforementioned schemes over 5000 implementation rounds. Throughout the simulations, the total number of nodes (700) in the proposed WSN remained operational up to 4500 rounds, with the attained throughput staying above 1000 (bit/sec). Notably, the number of dead nodes remains at 0 after 1500 rounds. Moreover, the probability of avoiding unsuitable cluster heads is the highest, and the packet delivery ratio consistently remains at or above 68%. Additionally, the energy consumption of nodes achieved by the GWO is the lowest and does not exceed 200 Joules, while the computational cost significantly declines to reach 30.5%.

The issue of DC to AC inverters has gained interest due to the growing use and investment in renewable energy sources as a competitive power source. This paper presents efficient technique for producing an AC sinusoidal signal from DC power. The primary objective is to create a highly effective inverter that generates a pure sinusoidal AC signal with the least amount of overall harmonic distortion. MATLAB Simulink was utilized to develop and simulate an inverter for this purpose. Four asymmetric DC voltage sources and sixteen power switches are included in the inverter design in a [ 1:2:7:11]Vdc ratio to create an inverter with 43 levels (-21 to 21).The value output voltage of the single-phase 43- Level (THD-V)equal 0.70%,while a three-phase system's value are 0.63% .The main contribution is the establishment and demonstration of a sound methodology and model based on multilevel inverter with high power and reduce number of switch.

Breast cancer incidence has been soaring sharply, and this is causing grave concern worldwide due to its high mortality rates. It should be correctly diagnosed in the early stages in order to achieve better patient outcomes. Over the last decade, there has been a great demand for diagnosis systems based on AI that could be used in breast cancer detection and classification. These computerized devices utilize deep learning algorithms to analyze medical scans thereby allowing for subtle abnormality recognition and distinguishing malignant from benign tumors. Computer-aided diagnosis named CAD systems can assist radiologists and pathologists to be more precise with their diagnoses while at the same time increasing productivity. Furthermore, recent advances in CNN architectures coupled with attention mechanisms have further improved CAD systems for breast cancer diagnosis. Attention-based CNN models focus on crucial regions hence enhancing classification accuracy and reliability. In this study, we introduce a new approach that improves the classification of breast cancer using GoogLeNet architecture modified by an attention mechanism that is based on image regions. The modified GoogLeNet has a spatial transformer network (STN), which allows it to focus on significant areas in breast histopathology images using selective attention. Through the attention mechanism, the model becomes better at learning discriminatory features that indicate different subtypes of breast cancer. In order to evaluate the effectiveness of this method, we implemented experiments using the BreaKHis dataset for classifying breast carcinomas. This dataset has been intentionally collected under various magnifications so as to facilitate binary and multiple classification tasks. These outcomes clearly show that modified GoogLeNet with attention outperforms the original GoogLeNet architecture in terms of accuracy. For binary classification, the proposed model demonstrated an accuracy rate of 98.08%, whereas GoogLeNet’s rate was 94.99%. For multi-class classification, at 100x magnification, this model achieved an accuracy of 94.63% while the accuracy of the original GoogLeNet was 85.06%. It is evident from these findings that the efficiency of breast cancer diagnosis significantly improved under this proposed approach. The findings of the study show that incorporating the modified GoogLeNet framework with an attention mechanism in CAD systems for breast cancer detection can improve their performance. Combining deep learning and attention models together can lead to more accurate treatment decisions and better patient results. More efforts are needed to further develop CAD systems in this area to assist ongoing endeavors towards upgrading them and ultimately contribute to saving many lives in the fight against breast cancer.

Bagging constitutes a prominent ensemble learning technique in contemporary machine learning. With this process, various instances of the base model are trained using various subsets of the training data that are extracted by bootstrapping. The resulting models are then aggregated, often through voting in a classification problem, to enhance performance and predictive power. Recent advances in bagging techniques include variants such as Random Forests, which introduce additional randomness by selecting a random subset of features in each partition and boosting algorithms that iteratively optimize the model's focus on misclassified instances. The efficacy of these strategies in enhancing the generality and adaptability of machine learning models has been impressive. There are many studies that confirm the ability of ensemble learning models to detect phishing attacks. However, the techniques used by these models that have enhanced their detection capabilities have not been highlighted. The study reached important results in terms of accuracy of up to 97% through the random forest model and the Particle swarm optimization algorithm. This study seeks to contribute to advancing the field of cybersecurity by providing a robust and interpretable machine learning-based classifier that can be integrated into a framework to detect phishing attacks by distinguishing between legitimate URLs and phishing URLs.

The English Tafseer Translation of the Holy Quran is essential for comprehending and interpreting Allah’s words for non-Arabic Muslims. This research adopted five different English translators (TR1-TR5) of chapter (Surah) Al-Baqarah and invested the advantages of the text clustering process to rank (prioritise) between these input five datasets. The absence of dataset ground truth (not standard datasets) requires the use of unsupervised learning (clustering technique) instead of other techniques (e.g. classification (supervised learning)). This study expanded the assessment to include both partitioning-based and hierarchical-based clustering algorithms. In a cluster based on partitioning, k-means is utilized. While for the hierarchical-based, the Agglomerative has been implemented. This research’s aim was achieved through a three-step procedure (stages). The first stage uses text cleansing to remove unnecessary words (Tokenisation, POS tagging, normalisation, stemming, and Stop-word removal). In addition, feature selection used VSM (Vector Space Model) and TF-IDF (Term Frequency-Inverse Document Frequency) to make the five corpora. The second stage implemented the clustering process. In the third stage, clustering validation was obtained using SC (Silhouette Coefficient) and DBI (Davies-Bouldin Index) metrics plus the execution time (ET). Principle Component Analysis (PCA) is used to visualise the clustering outputs. The results show, based on (ET, SC, and DBI) of the k-means algorithm, only ranks (1) and (3) demonstrate the same ranking for these five translators. In contrast, the Agglomerative algorithm shows the same five translators’ positions; each (ET, SC, and DBI) has a distinct rank. However, to obtain the optimal union rank, it is crucial to use a modern approach technique such as MCDM (Multi-Criteria Decision-making Analysis) in future work.

One of the most challenging goals in the fields of law enforcement and the court rooms is the ability to detect deception and false information, this is due to the major role it has on national security and the justice system. One of the most important ways recent literatures has explored to overcome this challenge, is detecting deception using artificial intelligence techniques to find patterns in verbal and nonverbal features. Text based analysis has been one of the most important modals for this task since written text can be found in transcribed audio, emails, online messaging services, news articles, and many more. In this work, a combination of machine learning techniques and data processing using hashing algorithms is used applied to n-gram feature representation on two of the largest datasets in deception detection field. Together, results of up to %94.59 accuracy were achieved. The paper reviews the most common techniques used in recent literature, it also details the methodology followed for data processing and model training to achieve these results.

This study proposes a lie detection module that leverages audio features in both the time and frequency domains to scrutinize key features and speech patterns for more precise dishonesty detection. The research focuses on addressing the limitations of traditional methods and suggests practical alternatives, aiming to contribute to the improvement of existing understanding and system architectures in the field of deception detection. where Detecting deception is critical in various fields, such as law enforcement, national security, and personal relationships. While traditional methods like polygraph exams are criticized for their reliability, emerging real-time technologies like voice stress analysis and speech processing techniques present alternatives. This research aims to enhance lie detection systems by exploring the potential of audio features, offering a more nuanced approach to identifying dishonesty, and overcoming the limitations of current methods. The study introduces a lie detection algorithm utilizing a real-world dataset collected with a handheld microphone, replicating authentic situations. Feature selection employs the random forest technique, focusing on the most significant features for training and testing. The algorithm achieved a 79% accuracy rate through rigorous examination, with Mel-frequency cepstral coefficients (MFCC) identified as the crucial feature for lie detection. This underscores the method's potential effectiveness in real-time fraud detection. However, further research is necessary to confirm its consistency across various datasets, situations, and demographic groups.

The rapid growth of big data and the emergence of new technologies have posed significant challenges in data compression, particularly in the domain of audio compression. Audio compression has gained widespread applications in advanced audio coding, MP3 encoding, web radio, and lossless audio coding techniques. In this paper, a newly proposed technique is present aimed at improving audio compression. The proposed technique consists of two main stages: preprocessing and the application of compression techniques. The preprocessing stage involves several steps. Firstly, the raw audio data is subjected to set normalization to ensure consistent scaling. Next, the data is divided into smaller segments to facilitate the application of Discrete Cosine Transform (DCT) and quantization on each segment. The quantization parameter is determined based on the bit per sample of the audio test file and the number of channels. To implement the Modified MLZW compression technique, the data is converted into byte values, and a word reference is constructed based on these bytes. To evaluate the performance of the proposed technique, multiple audio files have been utilized in experimental tests. The results demonstrate that the proposed technique achieves smaller MSE of 2.05*10-10 and significant improvements in terms of compression efficiency with reduction in processing time to 200 ms. The proposed methods works better and ensure that the audio quality is good and the system can be used for transmitting audio over different media that depend on lossy audio compression.