9 March 2021 in IEEE Access
Being one of the most widely used social media tools, Twitter is seen as an important source of information for acquiring people’s attitudes, emotions, views and feedbacks. Within this context, Twitter sentiment analysis techniques were developed to decide whether textual tweets express a positive or negative opinion. In contrast to lower classification performance of traditional algorithms, deep learning models, including Convolution Neural Network (CNN) and Bidirectional Long Short-Term Memory (Bi-LSTM), have achieved a significant result in sentiment analysis. Although CNN can extract high-level local features efficiently by using convolutional layer and max-pooling layer, it cannot effectively learn sequence of correlations. On the other hand, Bi-LSTM uses two LSTM directions to improve the contexts available to deep learning algorithms, but Bi-LSTM cannot extract local features in a parallel way. Therefore, applying a single CNN or single Bi-LSTM for sentiment analysis cannot achieve the optimal classification result. An integrating structure of CNN and Bi-LSTM model is proposed in this study. ConvBiLSTM is implemented; a word embedding model which converts tweets into numerical values, CNN layer receives feature embedding as input and produces smaller dimension of features, and the Bi-LSTM model takes the input from the CNN layer and produces classification result. Word2Vec and GloVe were distinctly applied to observe the impact of the word embedding result on the proposed model. ConvBiLSTM was applied with retrieved Tweets and SST-2 datasets. ConvBiLSTM model with Word2Vec on retrieved Tweets dataset outperformed the other models with 91.13% accuracy.
28 August 2023 in IEEE Access
Crime prevention relies on crime prediction as a crucial method to determine the most effective patrol strategy for law enforcement agencies. Various approaches and solutions have been utilized to predict criminal activity. Nonetheless the environment and nature of information for crime prediction is constantly changing. Although, potentially a useful source for gathering sentiments, social media content has been ignored by the prediction models. The utilization of social media for sharing information and ideas has experienced a significant surge. Twitter, in particular, is regarded as a valuable platform for gathering public sentiments, emotions, perspectives, and feedback. In this regard, techniques for analyzing the sentiment of tweets on Twitter have been developed to ascertain whether the textual content conveys a positive or negative viewpoint on crime incident. Therefore, our interest lies in investigating the potential and advantages of fusing the information of sentiment and crime modalities. In this paper, ConvBiLSTM is applied to train the model, features of both tweet and crime modalities were extracted independently at vector level and fused into a single representation that captures the information from all modalities. This study involved collecting and conducting experiments using two datasets. The first dataset consisted of crime incident data obtained from the Chicago police department, specifically covering the period between September 1 and September 30, 2019. The second dataset comprised tweets containing crime-related terminology specific to Chicago. The crime prediction using multimodal data fusion on ConvBiLSTM outperform against other models with 97.75% of accuracy.
12 February 2025 in EDRAAK
The advancements in deep learning (DL) and transfer learning (TL) have transformed artificial intelligence, particularly in image classification. This research examines the theoretical foundations of DL and TL, focusing on their applications in medical image classification, specifically distinguishing between COVID-19, viral pneumonia, and normal lung conditions. By leveraging GPU-enabled high-performance computing and large labeled datasets, DL models particularly Convolutional Neural Networks (CNNs) such as ResNet50 and VGG16 have achieved superior accuracy compared to traditional machine learning methods. This study explores feature extraction using pre-trained models, the implementation of classifiers like Support Vector Machines (SVM) and K-Nearest Neighbors (KNN), and the integration of multi-view learning techniques, such as early fusion. The results demonstrate the effectiveness of DL and TL in improving classification performance, highlighting their significant potential to advance global healthcare diagnostics.
10 January 2024 in Mesopotamian Journal of Arabic Language Studies
يفحص هذا التحليل الببليومتري تقدم البحث في اللغة العربية باستخدام مجموعة بيانات شاملة تم الحصول عليها من Scopus ومشاركتها على GitHub. تستخدم هذه الدراسة قياسات كمية لتحليل تقدم المجال، وتحديد الاتجاهات الهامة، والمساهمين البارزين، ومجالات التركيز المواضيعية. يُظهر التحليل وجود علاقة قوية بين الدراسات اللغوية والتحسينات التكنولوجية من خلال دمج مناهج الكمبيوتر مثل معالجة اللغة الطبيعية والتعلم العميق. تُظهر المساهمات الجغرافية اهتمامًا عالميًا بدراسات اللغة العربية، حيث تعد المملكة العربية السعودية وماليزيا والأردن وتونس الدول الأولى من حيث الإنتاج العلمي وتأثير الاستشهاد. وتتناول الدراسة دراسة تفصيلية للجوانب الاجتماعية اللغوية، مع التركيز بشكل خاص على النوع الاجتماعي، في أبحاث اللغة العربية. يتوسع التخصص بشكل ديناميكي ويظهر خصائص متعددة التخصصات من خلال النمو المستمر في مخرجات البحث ومجموعة متنوعة من المواضيع التي يشير إليها تحليل الكلمات الرئيسية. يوضح تحليل الشبكة المتزامنة الترابط بين قضايا البحث، حيث تعمل "اللغات العربية" كعقدة أساسية تربط بين العديد من التخصصات الفرعية. يوضح هذا...
8 March 2023 in Mesopotamian Journal of Computer Science
Accurate, timely, and safe administration of data from IoT devices is made possible by intelligent computing. As the number of IoT devices proliferates, more and more data will be collected, adding depth and breadth to the existing range of IoT services. Technology that incorporates entire systems on a single integrated circuit has improved to the point where more and more consumer electronics can support full-fledged operating systems. It is impractical to use a single computing model for the entire planet since doing so would cause severe network congestion and security holes. In order to solve this problem, we present a framework that utilizes both blockchain technology and edge computing to provide a lightweight data aggregation and secure data investigation solution for Internet of Things (IoT) applications without compromising data security or protection. We use a lightweight data aggregation process at the hub level to keep traffic flowing smoothly and quickly. In order to protect the safety of sensitive data and ensure the precision of IoT AI models, we also offer a multi-level fuzzy hashing process. Preliminary testing demonstrates the viability of the method presented here.