Metaverse insanlar için sanal dünya ile gerçek dünyayı bir araya getirir. Kullanıcıların avatarları ve hologramları kullanarak gerçek ve simüle edilmiş ortamlarda sorunsuz bir şekilde etkileşim kurmasını sağlar ve artırılmış ve sanal gerçeklik teknolojilerini kullanarak fiziksel dünyayı genişletme potansiyeline sahiptir. Sanal ortamlar ve sürükleyici oyunlar (Second Life, Fortnite, Roblox ve VRChat gibi) meta verinin öncülleri olarak tanımlanır ve kullanıcılarına tam, işlevsel ve kalıcı bir platform sunar. Eski bir kavram olmasına rağmen günümüzde yeni yeni kullanılmaya başlanmıştır. Bu çalışma kapsamında literatürde fazla örnekleri olmayan, yeni bir çalışma konusu olan metaverse terimi üzerinde durulmuştur. Metaverse’nin tanımı, kullanım alanları, önemi, faydaları ve zorlukları anlatılmış ve ileride bu teknolojiyle ile neler yapılabileceği ile alakalı bilgiler verilmiştir. Bu teknoloji tanıtılarak ileride yapılacak çalışmalar için literatüre katkı sağlanması amaçlanmıştır. Ayrıca Metaverse’de kullanılan teknolojiler tanıtılmıştır ve konu detaylı bir şekilde ele alınmıştır. Bu çalışma bu alanda çalışma yapacak araştırmacılar için örnek teşkil edecektir.

Healthy lettuce receives all the necessary nutrients, such as calcium, magnesium, nitrogen, phosphorus, and potassium for optimum growth. Deficiency of one of these vitamins causes growth retardation in plants. Vitamin deficiencies in lettuce grown in hydroponic systems can negatively affect plant development and create various problems. Therefore, it is critical to provide all the vitamin and mineral balances that plants need in hydroponic systems. In this study, the development of lettuce, which has an important place in daily life, was examined in detail. Deep learning techniques were applied to images of lettuce plants using various You Only Look Once (YOLO) models (YOLOV12 and YOLOV11). Various metrics were calculated for each YOLO version to evaluate performance. These metrics include precision, recall, mean average precision (mAP)50 (average accuracy), and mAP50–95 (average accuracy at different intersection over union [IoU]), thresholds). A balance was created between the classes on the lettuce images obtained using data augmentation techniques, which allowed the model to be trained more robustly. In the tests conducted with the augmented dataset, the mAP50 value of the YOLOV12 model was calculated as 0.984 and the mAP50–95 value as 0.957. This reveals the high performance of the model in correctly recognizing nutritional deficiencies. On the other hand, the YOLOV11 model achieved lower results, such as 0.975 mAP50 and 0.944 mAP50–95 in all classes. This study significantly improved the object detection performance by integrating novel attention modules into YOLOV11 and YOLOV12 models to detect nutrient deficiencies of lettuce plants early and accurately in soilless agriculture. Not only the comparison of models, but also the generalization and robustness of the model were strengthened with innovative modules that provide dynamic adaptation to environmental variables. This novel approach provides tangible contributions to both academic literature and practical applications in AI‐supported early detection and yield optimization in sustainable hydroponic agriculture.

Collaborative Filtering or recommender systems use a database for new users and new items about theirs preferences. It is very important to make private suggestions to users, keep their interest alive with admirable suggestions. Collaborative Filtering (CF) is a commonly used system to meet this end. However, despite the fact that CF systems are widely used, traditional CF techniques are unable to track the preferences of users over a period of time. For this reason, “temporal dynamics” has become an important notion in recommendation systems. In this study a new method has been employed to provide customized suggestions to users whose tastes may have changed over time. The proposed system is different from the traditional user-based CF technique and item-based CF technique in that it examines the dates users ranked products and uses this data to help determine user preferences. The evaluation process has been performed on Netflix data in order to measure the success of the system and compare the results with traditional user-based CF technique and item-based C technique. The results are encouraging and the quality of the predictions were significantly improved.

In several application, emotion recognition from the speech signal has been research topic since many years. To determine the emotions from the speech signal, many systems have been developed. To solve the speaker emotion recognition problem, hybrid model is proposed to classify five speech emotions, including anger, sadness, fear, happiness and neutral. The aim this study of was to actualize automatic voice and speech emotion recognition system using hybrid model taking Turkish sound forms and properties into consideration. Approximately 3000 Turkish voice samples of words and clauses with differing lengths have been collected from 25 males and 25 females. In this study, an authentic and unique Turkish database has been used. Features of these voice samples have been obtained using Mel Frequency Cepstral Coefficients (MFCC) and Mel Frequency Discrete Wavelet Coefficients (MFDWC). Moreover, spectral features of these voice samples have been obtained using Support Vector Machine (SVM). Feature vectors of the voice samples obtained have been trained with such methods as Gauss Mixture Model( GMM), Artifical Neural Network (ANN), Dynamic Time Warping (DTW), Hidden Markov Model (HMM) and hybrid model(GMM with combined SVM). This hybrid model has been carried out by combining with SVM and GMM. In first stage of this model, with SVM has been performed subsets obtained vector of spectral features. In the second phase, a set of training and tests have been formed from these spectral features. In the test phase, owner of a given voice sample has been identified taking the trained voice samples into consideration. Results and performances of the algorithms employed in the study for classification have been also demonstrated in a comparative manner.

Access control models are an important tool developed for securing today’s data systems. Institutions use the access control models specifically to define who their employees are, what they can do, which resources they can reach, and which processes they can perform and use them to manage the whole process. This is a very hard and costly process for institutions with distributed database systems. However, access control models cannot be implemented in a qualified way due to the fact that the conditions for defining users’ demands to reach resources distributed on different servers, one of which is consequentially bound to the other, the verification and authorization of those user demands, and being able to monitor the actions of the users cannot be configured in an efficient way all the time. With our model suggested in this study, the aim is to automatically calculate the permissions and access levels of all users defined in the distributed database systems for the objects, and, in this way, we will reach a more efficient decision as to which objects the users can access while preventing their access to the information they do not need. Our proposed model in this study has been applied to real life data clusters from organizations providing health and education services and a public service. With the proposed model, all models have been run on servers sharing resources in a private network. The performance of the proposed model has been compared to that of traditional access models. It was confirmed that the proposed model presented an access control model providing more accurate access level results as well as being scalable to many distributed database systems.

Although cancer is a broad term, it is an important problem that causes a high rate of death. It indicates disease that occurs when cellular changes cause cells to grow and divide uncontrollably. Cancerous cells can form tumors, damage the immune system, and cause other aberrations that prevent the body from functioning properly. In particular, lung cancer is the type of cancer with the highest death rate in the last five years. The important thing in lung cancer is early diagnosis and diagnosis to ensure the survival of patients. In this study, a model that automatically detects lung cancer has been developed in order to detect lung cancer early and take the necessary precautions at the first stage of the disease. The aim of our study is to diagnose the presence of lung cancer cells based on attributes and information from human symptoms. In our study, a dataset consisting of 26 features and 3 classes determining lung cancer as low, medium and high was used. Some preprocessing and transformations have been done to make the data more suitable for predictive analysis. The first 26 features were used as the inputs of the model. The lung cancer feature was used as the predicted output based on the input features. The values of the attributes are normalized. In the first stage, data preprocessing was carried out with various data mining techniques in order to process the data more easily and increase the performance. In the proposed model, feedforward backpropagation Artificial Neural Network model was used to detect the presence of lung cancer in the body of the person. Unlike other studies, the factors that cause lung cancer by producing our own neural network have been determined. When the results obtained are compared with the studies in the literature, it has been observed that it gives very successful results.