31 December 2023 in Yüzüncü Yıl Üniversitesi Tarım Bilimleri Dergisi
Agriculture has a big impact on society because it is essential for a large percentage of our food. The issue of hunger is getting worse by a growing population in many nations, resulting in food shortages or insufficiencies. To meet the world's food needs, it is ever more crucial to provide crop protection, conduct detailed land surveys, and predict crop yields. To calculate the estimated number of crops that are produced in a year, this research focuses on the use of machine learning techniques to predict crop yield and recommend crops with the highest yield in the Northeast region of India. The crop market's fluctuations in prices may be controlled with the aid of this information. To estimate agricultural crop yields, this study accurately evaluates a range of machine learning regression models, such as Linear Regression, Decision Tree, Random Forest, Gradient Boosting, XGBoost (eXtreme Gradient Boosting), and AdaBoost. With a 0.98 R2 score for the XGBoost and 0.96 for the Random Forest, they performed better than the other models.
30 June 2024 in Yüzüncü Yıl Üniversitesi Tarım Bilimleri Dergisi
In this research, the primary objective is to tackle the pressing issue of identifying and effectively managing diseases in rice plants, a problem that can results in substantial crop losses and pose a severe threat to food security. The study employs Convolutional Neural Networks (CNNs), a type of deep learning model widely used for image analysis, to conduct an extensive investigation using a sizable dataset comprising 5,932 RGB images. These images represent four distinct disease classes in rice plants: Bacterial Leaf Blight (BLB), Blast, Brownspot, and Tungro. To conduct this research, the dataset is split into two subsets: a training set, which comprises 80% of the data, and a testing set, which makes up the remaining 20%. This division allows for a systematic evaluation of the performance of four different CNN architectures: VGGNet, ResNet, MobileNet, and a simpler CNN model. The results of this study consistently show that ResNet and MobileNet outperform the other CNN architectures in terms of their ability to accurately detect diseases in rice plants. These two models consistently achieve remarkable accuracy in identifying these diseases. The research findings not only emphasize the potential of deep learning techniques in addressing the critical issue of rice crop diseases but also highlights the significant role that ResNet and MobileNet play in strengthening crop protection efforts and contributing to global food security.
24 January 2026 in Journal of Information & Knowledge Management
10 June 2026 in Journal of Automation, Mobile Robotics and Intelligent Systems
In recommender systems, collaborative filtering (CF) is a crucial technique, but it often struggles with the data sparsity issue, which impact the recommendation accuracy. To address this challenge we have proposed a Co-training Ensemble Learning (CTEL) technique that integrates item-based Collaborative Filtering, user-based collaborative filtering (CF), and Singular Value Decomposition (SVD) through a structured stacking methodology to improve the recommendation performance. The co-training procedure, which creates pseudo-labels for unlabeled data based on a confidence threshold, is used to iteratively improve the user-based and item-based CF models after they have been originally trained. These models produce predictions for validation and test sets, in conjunction with the independently trained SVD model. These forecasts yield meta-features, which include extra statistical variables like variance and product of predictions. The Linear Regression model is trained as the meta-learner to find the prediction of the base models in the best possible way using K-Fold cross-validation. Mean Absolute Error (MAE), Mean Squared Error (MSE), and Root Mean Squared Error (RMSE) are used to assess the final model's performance on a test set. The outcomes confirm the effectiveness of the co-training and stacking strategy by showing notable increases in prediction accuracy when compared with existing techniques. Using the advantages of collaborative filtering approaches and matrix-based approach, the proposed model offers a comprehensive foundation for creating advanced recommendation systems.
31 March 2026 in Gazi University Journal of Science Part A: Engineering and Innovation
The proposed research presents a novel method of enhancing credit-card fraud detection using Secure Multi-Party Computation (SMPC), a privacy-saving method. At present data privacy is of utmost concern and the proposed methodology can eliminate the problem of carrying out fruitful fraud detection without putting privacy of user data in danger. Using SMPC, several entities are able to work with transaction data in a collaborative fashion without revealing any of the individual data, establishing privacy compliance. The study aims at a privacy-preserving framework for identifying fraud in credit card payments by linking SMPC and the FP-Growth algorithm. The usage of SMPC means that cooperative calculations may be done without revealing sensitive data, avoiding the possibility of data leakage. As a powerful Association Rule Mining (ARM) approach, FP-Growth protects the secrecy of individual transactions while effectively detecting hidden patterns and correlations inside transaction datasets. The framework is tested using extensive performance metrics including as accuracy, precision, recall, and F1-score, which serve as a robust standard for determining its dependability. The results show that the suggested technique is highly effective in identifying fraudulent actions while ensuring tight privacy protection.