E-businesses often face challenges related to customer service and communication, leading to increased dissatisfaction among customers and potential damage to the brand. To address these challenges, data-driven and AI-based approaches have emerged, including predictive analytics for optimizing customer interactions and chatbots powered by AI and NLP technologies. This study focuses on developing a hybrid rule-based and extractive-based chatbot for e-business, which can handle both routine and complex inquiries, ensuring quick and accurate responses to improve communication problems. The rule-based QA method used in the chatbot demonstrated high precision and accuracy in providing answers to user queries. The rule-based approach achieved impressive 98% accuracy and 97% precision rates among 1684 queries. The extractive-based approach received positive feedback, with 91% of users rating it as “good” or “excellent” and an average user satisfaction score of 4.38. General user satisfaction was notably high, with an average Likert score of 4.29, and 54% of participants gave the highest score of 5. Communication time was significantly improved, as the chatbot reduced average response times to 41 s, compared to the previous 20-min average for inquiries.

As a result of global epidemics and threats, higher education began to widely use e-learning platforms. With the proliferation of online learning in higher education, a systematic and rigorous approach is needed for universities to make informed decisions when selecting e-learning platforms. Presenting a Pythagorean fuzzy hybrid multi-criteria decision-making strategy that aids in solving this challenge is the study's goal. This study presents a comprehensive analysis of four prominent e-learning platforms - Google Meet, Microsoft Teams, Skype, and Zoom - utilizing the Pythagorean Fuzzy Analytic Hierarchy Process (AHP) and Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) methodologies. By structuring and prioritizing a range of criteria using the Pythagorean Fuzzy AHP model, the study establishes a well-defined framework for the assessment process. Subsequently, the Pythagorean Fuzzy TOPSIS method is applied to rank these platforms based on their overall performance against the identified criteria. The outcomes of this research enable universities to tailor their e-learning platform selection to the unique requirements of their educational programs, ultimately promoting enhanced engagement, accessibility, and learning outcomes for both students and faculty.

International Diabetes Federation (IDF) reports that diabetes is a rapidly growing illness. About 463 million adults between 20-79 years have diabetes. There are also millions of undiagnosed patients. It is estimated that there will be about 578 million diabetics by 2030 [1]. Diabetes reasons different eye diseases. Diabetic retinopathy (DR) is one of them and is also one of the most common vision loss or blindness worldwide. DR progresses slowly and has few indicators in the early stages. It makes the diagnosis of DR a problematic task. Automated systems promise to support the diagnosis of DR. Many deep learning-based models have been developed for DR classification. This study aims to support ophthalmologists in the diagnosis process and increase the diagnosis performance of DR through a hybrid model. A publicly available Messidor-2 dataset was used in this study, comprised of retinal images. In the proposed model, images were pre-processed, and a deep learning model, namely, InceptionV3, was used in feature extraction, where a transfer learning approach is applied. Next, the number of features in obtained feature vectors was decreased with feature selection by Simulated Annealing. Lastly, the best representation features were used in the XGBoost model. The XGBoost algorithm gives an accuracy of 92.55% in a binary classification task. This study shows that a pre-trained ConvNet with a metaheuristic algorithm for feature selection gives a satisfactory result in the diagnosis of DR.

İnternetin, Dünya’da iletişim ve bilgiye anında erişimde insanlık tarihinin en önemli keşfi olduğu yadsınamaz bir gerçekliktir. Günümüzde internetin tüm Dünya’da yaygın bir teknoloji olmasında kullanılan araç, fiber optik kablolardır. Fiber optik kabloların bu bakımdan önemi, üretiminin nasıl gerçekleştirilmekte olduğu merakını doğurmaktadır. Fiber optik kablo üretiminde, özellikle fiberlerin zayıflama değerleri kablonun optik özellikleri açısından önemli bir faktördür. Zayıflama değerleri belirli bir değerin üzerinde olan kablolar atıl ürün olarak üretimden ayrılmaktadır. Bu durum işletme verimliliğini olumsuz etkilemektedir. Bu çalışma, Türkiye’nin en önemli fiber optik kablo üreticilerinden birinde fiber optik kablo üretimine ait özellikleri incelemektedir ve Taguchi deney tasarımı yöntemiyle en uygun çalışma faktörlerine karar vermektedir. Fiber optik kablonun üretiminde zayıflamaya etkisi düşünülen çalışma faktörleri beyin fırtınası tekniği ile ilgili uzman ve akademisyenlerden oluşan bir ekip tarafından belirlenmiştir. Optimal çalışma koşullarını belirlemek için L27 (37) ortogonal matris kullanılmış ve gerçekleştirilen deneyler ile tespit edilen parametrelerin zayıflama değeri üzerindeki etkisi incelenmiştir. Ayrıca her proses parametresinin zayıflama üzerindeki önemi Varyans Analizi (ANOVA) testi ile belirlenmiştir. Analiz sonuçları tartışılmış ve şirket için optimum çalışma koşulları, 18 °C Ortam Sıcaklığı, 3,0-6,0 mm Excess Uzunluğu, Orta kalite PBT, Yüksek Kalite Thixojel, 80 mm Hatve Uzunluğu, 185-190 °C Extruder Sıcaklığı, 1,0-1,5 bar Çekme Kuvveti olarak belirlenmiştir.