الأهداف: تُعد الثقة في اتخاذ قرارات الاستثمار من الأدوات الرئيسية لتقييم أسواق الأسهم. يتوجب على كل مستثمر تحديد أسعار الأسهم المستقبلية قبل تطوير استراتيجيات الاستثمار لتعزيز عائداتهم. ومع ذلك، فإن الطبيعة غير الخطية والمعقدة والمتقلبة والديناميكية لبيانات الأسهم تجعل من التنبؤ بأسعار الأسهم مهمة صعبة. تهدف هذه الدراسة إلى مواجهة هذه التحديات وتحسين دقة التنبؤ باستخدام تقنيات النمذجة المتقدمة. المنهجية: اقترحنا نموذجًا هجينًا للشبكات العصبية يدمج بين وحدات التكرار المغلقة ثنائية الاتجاه (BiGRU) والشبكات العصبية الالتفافية أحادية البعد (CNNs). تم إجراء تجارب على خمس مجموعات بيانات لأسعار الأسهم، تضمنت ثلاث أدوات أسهم فردية ومؤشرين للأداء في أسواق الأسهم. شملت معايير التقييم متوسط الخطأ المطلق (MAE) وجذر متوسط مربع الخطأ (RMSE). النتائج: حقق النموذج الهجين المقترح أدنى قيم للـ MAE لمجموعات البيانات الفردية، حيث تراوحت بين 0.357 و1.747. كما أظهرت التجارب انخفاضًا يتراوح بين 42% و91% في قيم RMSE مقارنةً بالأساليب القياسية المعتمدة على CNN وGRU، مما يعكس تحسنًا كبيرًا في دقة التنبؤ. الأبعاد: تشير النتائج إلى أن النموذج الهجين القائم على...

Hazard recognition and prevention techniques utilize augmented reality (AR) and virtual reality (VR) technology for immersive training and identification of potential hazards. Currently, no scoping reviews exist specifically on hazard recognition and prevention using AR and VR, which is crucial for forming evidence-based guidelines. This scoping review aimed to identify the use of AR and VR in hazard recognition and prevention, identify research gaps, and provide future recommendations. This Joanna Brings Institute (JBI) approach was used to investigate hazard recognition and prevention with AR and VR interventions. A total of six reliable databases were included: Google Scholar, IEEE Xplore, ACM Digital Library, PubMed, Scopus, and APA PsycNet, focusing on records and publications from the last five years. The review identified 5438 publications, with 22 meeting final inclusion criteria. The major findings include the use of AR and VR for teaching new skills, ensuring health and safety, enhancing realism, developing cognitive skills, improving performance, increasing engagement, and stimulating a sense of presence. Limitations identified include limited or non-diverse sample size, absence of real-world scenarios or contextual learning, inadequate game mechanics, insufficient validation of test results, lack of dynamic game behaviour, insufficient expert inclusion, technology issues, long-term impact assessment, and cost comparisons with traditional approaches. Technological issues have emerged as an additional impediment, limiting the efficacy of AR/VR applications in hazard recognition and prevention. In light of these findings, future research should focus on integrating contextual learning, expert inclusion, sample recruitment, expert validation, long-term impact assessment, and resolving technological challenges to enhance AR/VR applications’ efficacy and real-world viability.

Boolean expressions are widely used to model decisions or conditions of a specification or source program. The MUMCUT, which is designed to detect seven common faults where Boolean expressions under test are assumed to be in Irredundant Disjunctive Normal Form (IDNF), is an efficient fault-based test case selection strategy in terms of the fault-detection capacity and the size of selected test suite. Following up our previous work that reported the fault-detection capacity of the MUMCUT when it is applied to general form Boolean expressions, in this paper we present the characteristic of the types of single faults committed in general Boolean expressions that a MUMCUT test suite fails to detect, analyze the certainty why a MUMCUT test suite fails to detect these types of undetected faults, and provide some extensions to enhance the detection capacity of the MUMCUT for these types of undetected faults.

A common practice to capture the non-stationary characteristics of the time series data in Artificial Neural Network (ANN) is by randomly dividing the whole set of available data into training, validation and testing, i.e. the data in validation and testing are represented in the training data. Consequently, the usability of the developed model on data not represented by the training data used during the network model development process is always doubtful. In this work, we present a back-propagation neural network (BNN) model trained using one-day history data to predict soil moisture data at 1 km resolution for two future dates. Specifically, high soil moisture values were observed in the training data while the testing data were characterized by drier conditions due to minimal or no rainfall. Our model uses separate mean and standard deviation statistics values from the training and testing data, respectively, to the z-normalized data. With data pre-processed using this method, the BNN model next uses a moving window of size 4 km × 4 km to capture the spatial variability of the soil moisture throughout the 40 km × 40 km study area. The coupling of the normalization and moving window method managed to achieve average soil moisture with Root Mean Square (RMSE) of 3.67% and correlation coefficient,R²of 0.89. By only using the suggested normalization without the moving window method, the BNN model managed to achieve an average RMSE of barely 5.82% withR² = 0.83. When comparing with the normal practice of using the same mean and standard deviation statistics of the training data in the testing data, the retrieval accuracy of the BNN model deteriorates to 8.86% withR² = 0.32. The experiment results demonstrated that the proposed coupling method performed better in terms of both RMSE and R²for soil moisture retrieval.