Journal of Computers and Intelligent Systems

Optimizing Crop Yield Forecasts Using Quantum Machine Learning Techniques with High-Dimensional Soil and Weather Data

Jamshaid Basit — 2025-03-23

This paper focuses solely on the possibility of applying quantum machine learning methods to increase crop yield prediction accuracy based on multi-feature soil and climate data. The main goal is to increase the efficiency of crop yield prediction models, which are critical for increasing a nation's production and food ratio. Complexity also throws off supervised analytical methods, and nonlinearity grew as the agricultural industry expanded its fields. These fields now encompass a wider range of interconnected elements, including soil type and nutrient content, their relationship to soil water content, air temperature, rainfall, and other factors. In this research, we use quantum computing to solve the problem of handling high-order data more proficiently than the same problems formulated in classical computers. In this paper, we developed and incorporated QSVM and QNN into conventional machine learning models to learn from large and highly complex datasets containing multiple years' worth of regional and temporal information on soil and weather. We believe these models can reveal patterns that QSVM and QNN are better equipped to detect due to their scalability and ability to compute over large datasets. As a result, the quantum-enhanced models outperform the conventional methods in terms of predictive power, demonstrating superior MSE values and robustness values. Specifically, the integration of quantum techniques enhanced the generalization ability because of the highly nonlinear relationship between the variables. These results suggest that QML could significantly improve crop yield estimates, as its predictions are more accurate and directly applicable to agricultural practices and policies. This study would expand the literature on the application of quantum computing in agriculture because it is an emerging field that holds potential for addressing various challenges in food production. In the domain of crop yield prediction, we are laying down the foundations for less vulnerable farming structures that are able to meet the future climate conditions and the growing global food requirements. Thus, the study calls for more research on potential quantum-based solutions in other essential use cases in agriculture.

Adaptive YOLO-V8 for Low-Earth Orbit Debris Detection Improving Detection Accuracy with Dynamic Training

Hira Arshad — 2025-02-25

It is crucial to identify methods to detect debris due to a new tendency that is emerging in LEO orbits, which is threatening the functionality of satellites and interplanetary missions. This research solves this problem using an improved YOLO-V8 model that enables the detection of space debris with higher precision while using adaptive dynamic learning approaches. It was critical for our model to be able to identify and categorize as many kinds of objects as possible, and our database currently contains 11 object classes, including space debris and satellites. To specifically detect small and moving objects, we utilized the YOLO-V8 model, tailoring the train options to the unique object detections in this class. For the training of our model, we used a large quantity of data in addition to the images from these 11 classes and used SGD as our optimizer with the learning rate of 0.25 and individual weight decay parameters. Additionally, we utilized blur and grayscale transforms to enhance the model through data augmentation. By comparing the obtained results, we can observe enhanced detection accuracy in each class separately, as well as a general boost in prediction and recall. Due to the cross-entropy function's flexibility, the model was able to perform well on various object sizes and speeds in an orbital context, making detection consistent. A lot of fine tuning was required in the training parameters in order to get the desired or even better results devoid of false positive detection. This paper describes how YOLO-V8 with adaptive training achieved outstanding results for object and debris detection in low Earth orbit (LEO) to improve space usage safety and define better approaches to space debris management.

Comparative Analysis of Deep Learning Architectures for Customer Churn Prediction in the Banking Sector

Azmeena Sheikh — 2024-03-23

Customers’ churn is a critical problem for the banking industry since the prediction of customers’ attrition impacts the business’s key outcomes and decision-making. As such, the research aim of this study is to assess the performance of different deep learning networks in the context of customer churn rate estimation with banking datasets. This study specifically compares feedforward neural networks, long short-term memory networks, convolutional neural networks, and multi-layer perceptrons. We train and test the models on a data set that includes customer details such as credit score, age, balance, and tenure. Hence, we measure each model’s ability depending on factors like accuracy, precision, recall, F1 score, and the ROC-AUC. However, for the evaluation of the models, we use various visualizations, including confusion matrix, receiving operating characteristic curve, precision-recall curve, and learning curves. The results show that all models can achieve comparable performance, but there are some models with specific edges. For example, long short-term memory networks, a type of RNN, excel at modeling sequential relationships in the data, whereas convolutional neural networks craft intricate structures within the data input. This work's main ideas include examining the characteristics and potential of various deep learning designs in the context of customer churn prediction while comparing the architectures. The primary goal of this study is to analyze and identify the most effective deep learning model for customer churn prediction, as well as provide recommendations for banks to improve customer retention. Thus, the results emphasize the importance of selecting an adequate model based on the data's characteristics and prediction goals. It is in this vein that the study proposes to advance the understanding of the deep learning models above with a view to informing banking institutions on how best to address customer churn problems.

Performance of Deep Learning in Malware Classification

humza rana — 2025-03-23

The malware program is designed to harm the user's data and information. With the new advancements in technology, all the business systems now get through a network. Large-scale businesses are also now in online systems. The security of these systems is essential. The malware programmer developed an advanced type of code that breaks the user security, user information, and money. The malware is of different types. The deep learning is used to classify the modern type of malware. In this survey, the deep learning models that have been used in classifying the malware are studied. Their model performance, their datasets, and preprocessing are discussed in this paper. After an overview of the deep learning model in malware classification with their preprocessing and datasets, we discuss further research direction, to improve the security.

Detection of Phishing Attack by using LightGBM&Xgbost

Ansar Munir Shah — 2025-03-27

Phishing attacks provide a significant security risk to both individuals and organizations. To steal sensitive information, these assaults are typically carried out by creating phony websites that substantially resemble actual ones. This research employs these phishing attacks, by using AI techniques that have lately been used to look at the URLs of these phony websites. We have proposed the increasing sophistication and frequency of phishing attacks, highlighting the need for an enhanced AI-based model to detect such attacks effectively. Involves LightGBM, Xgbost, and using a hybrid model of a LightGBM, Xgboost classifier to train and test data for detecting phishing attacks on URLs. There are several feature extraction techniques used to detect URL phishing attacks. To identify if a website is a phishing assault or not, these attributes are then given to a LightGBM and Xgboost classifier. As compared to the previous research model’s accuracy was 93%, Hence The current proposed results of combining training and testing datasets on LightGBM and XgBoost give a 96% accuracy and improve the quality-of-evaluation metrics of the feature of the URLs to detecting phishing attack detection.

A Comparative Study of Traditional and Hybrid Models for Text Classification

Muhammad Sabir — 2025-03-29

Natural Language Processing (NLP) is a fundamental task that is essential for the automation of the categorization of textual data using an existing set of categories, such as sentiment analysis, spam detection, fake news detection, etc. Due to the interpretability and also efficiency, the Logistic Regression (LR), Support Vector Machine (SVM), and Random Forest (RF) have been very popularly used for text classification under traditional machine learning models. Yet, such models fail in modeling contextual linkages and semantic subtleties as they would be necessary to handle text with complex structure. As such, hybrid models that couple the two traditional and deep learning techniques have emerged as a potential way to address these problems.

In the note, I review all efforts of text classification that have the potentials of contributing to my classification task, which includes traditional machine learning models, hybrid models, and deep learning models. The AG News dataset is used for evaluation and accuracy, precision, recall and F1 score are used to measure the performance of the models. Finally, the results suggest that both deep learning based hybrid models such as BERT + SVM Hybrid Model (95.7%) and CNN + LSTM Hybrid Model (94.5%) surpass the performance of any traditional or ensemble learning based models by the exploitation of contextual embeddings and sequential modeling. XGBoost (92.8% accuracy) and Bagging Classifier (91.5% accuracy) of ensemble learning models have good generalization as well as stability compared to standalone learner.

Though the hybrid models offer superior classification performance at the sacrifice of computational resources, longer training times, there are tradeoffs in regards to the model classes and the problem. It brings out the tradeoffs made by traditional, ensemble, and the deep learning based hybrid models toward the applicability of the same towards different classification of text. The findings establish a platform towards choosing the best suitable classification model under performance requirements and computational constraints for researchers and practitioners.