International Journal of Artificial Intelligence & Mathematical Sciences

Global dynamics of a staged progression model for whooping cough

Samuel Tosin Akinyemi — 2025-12-03

This paper introduces a comprehensive stability analysis of a pertussis (whooping cough) model that incorporates staged disease progression. To examine the asymptotic and symptomatic behavior of the disease with respect the model’s equilibria, a qualitative analysis of the model is conducted. The local stability of the disease-free equilibrium is demonstrated using the Jacobian stability method, demonstrating its local asymptotic stability. Additionally, the comparison method is utilized to demonstrate the global stability of the model, establishing that the disease-free equilibrium achieves global asymptotic stability when the basic reproduction number falls below one. Numerical simulations based on baseline data further validate the analytical results. Furthermore, the research explores the influence of varying some of the model parameters.

Enhancing Agricultural Efficiency with Uav-Aided Smart Irrigation Systems

Samaila Umaru — 2025-12-04

Agriculture faces a critical obstacle: the limited availability of timely, precise data needed for effective decision-making and farm
operations. To maximize production, minimize expenses, and increase yields, farmers need immediate access to actionable
information. Technology-driven crop management approaches, known as precision agriculture, present transformative opportunities
for agricultural stakeholders. Among these technological advances, unmanned aerial vehicles (UAVs) have emerged as innovative
instruments that deliver multiple advantages. When outfitted with sophisticated sensors, these aerial platforms can acquire high
resolution imagery, track biological and environmental stress factors, identify pest infestations and plant diseases, and support
targeted spraying and pollination activities. UAV applications extend to monitoring livestock, tracking natural resources, and various
other functions. Through UAV deployment, farmers obtain vital field and environmental data at significantly lower costs compared
to conventional approaches. The substantial data volumes produced by UAVs enable analytical processes that yield practical
recommendations, resulting in heightened agricultural output, minimized resource waste, and improved environmental stewardship.
This study examines the contemporary application of UAVs in farming contexts, intelligent irrigation technologies, and opportunities
for their combined implementation to advance agricultural efficiency and ecological sustainability.

Machine Learning-based Smart Students' Complaint Resolution System

Rimsha Javed — 2025-12-03

For enhancing the quality of any educational institution, student feedback and complaints play an important role. Effectively handling complaints is crucial for ensuring student happiness and expediting the resolution of issues. Traditional manual complaint handling processes, however, are frequently cumbersome and ineffective, resulting in delays that irritate employees and students alike. Our research presents an academic facilitation system that automatically categorizes student complaints by department and particular aspect to address these issues. Our study focuses only on complaints pertaining to the academic setting, in contrast to many other approaches that use supervised and unsupervised learning techniques for common complaint management. We are manually an-notated the data across four departments and forty-two expectations after conducting a survey to gather actual student input because there was no previous dataset of academic complaints accessible. Afterwards, machine learning and deep learning models for classification and aspect identification were trained using this annotated dataset. Bidirectional Encoder Representations from Transformers (BERT), Convolutional Neural Networks (CNN), Support Vector Machines (SVM), Random Forest (RF), Decision Trees (DT), and One Vs All classifiers were among the several algorithms we tested with. Standard performance measures were used to evaluate the models. Among these, the Random Forest model performed well, with 65% accuracy across all our test data, and the One Vs All classifier reached 94% accuracy in both department classification and aspect identification. The study did have several drawbacks, though. The medium-length complaint letters that made up our dataset did not specifically address latent meanings. In the future work, we intend to increase the size of the dataset, incorporate longer complaint texts, and investigate cutting-edge methods for identifying hidden or implicit meanings in student complaints.

A Hybrid Approach Combining LSTM Prediction, Genetic Algorithm Optimization, and Reinforcement Learning for Adaptive Traffic Signal Management

Munir Ahmed — 2025-12-03

The dynamically growing urban population and the number of vehicles have exacerbated traffic congestion; thus, intelligent and scalable traffic management systems have to be developed. This study will present an AI-based optimization system, which is based on the parallel computing, to be used in optimizing real-time traffic prediction, route, and signal control. The structure of the system consists of three layers linked to each other: the data collection layer, which processes traffic, weather, and accident data used by various sources; an AI processing layer, integrating models of deep learning and optimization models; and a decision layer reporting the results of the application to both traffic authorities and commuters. LSTM networks are deployed to predict traffic flow estimating the temporal relationships and forecasting congestion under different conditions. GA and RL are used in the optimization of routes and adaptive traffic signal, respectively, thereby means of efficient vehicle routes and shorter delays. The integration with distributed data processing engine (Apache Spark) and deep learning frameworks accelerated by GPU (TensorFlow / PyTorch) will be used to guarantee scalability and real-time performance. Validation was done experimentally using the SUMO simulation platform, which has proven effectiveness of the proposed framework. The results of the LSTM model demonstrated a much lower level of errors in the prediction (MAE = 0.23, RMSE = 0.28) than in its traditional counterparts. It was found that the RL-based signal controller yielded an average waiting time 28-percent lower per intersection and 18-percent higher throughput when compared to the fixed-time scheduling method. Moreover, parallel processing experiments indicated that the average data processing latency decreased when scaling Spark clusters between 4 and 16 nodes (reduction of 2.1s to 0.5s), which means it is robust and scalable.

Deep Learning for Precision Agriculture: Tomato Leaf Disease Diagnosis Using Convolutional Neural Networks

Sajjad Ahmed — 2025-12-04

One of the large crops in Pakistan is tomato, which is cultivated on a significant area and yields approximately 0.31 million tons every year. It is a major income earner to agricultural communities, and it is a key contributor to nutritional requirements. Though not underestimated, diseases and unfriendly environmental conditions often lower tomatoes yields particularly at the initial stages of growth. The current research proposes a Convolutional Neural Network (CNN) architectural model of detecting and classifying significant tomato leaf diseases. The methodology involves taking of images, preparation of data, preprocessing, training of CNN, and measurement of performance by conventional accuracy and precision measures. Empirical evidence has shown that the model had a training and validation accuracies of about 90 and 85-88 respectively, showing its efficiency in the detection of diseased and healthy leaves. This suggested framework provides a feasible input to the improvement of the disease management system to enhance sustainable tomato production and to empower the lives of farmers in Pakistan.

FASHION ACCESSORY RECOMMENDATION SYSTEM

Sundus Latif — 2025-12-04

This research focuses on the development of an AI-powered system that helps people choose accessories that match their
outfit and enhance outfit coordination by analyzing the dominant color in the fabric image. The system uses advanced color
extraction to analyze the main shades in fabric and generates accessory suggestions that complement with the analyzed
colors. It also integrates current fashion trend insights, such as geometric patterns and earthly tones, to make stylish and
trendy recommendations suitable for fashion designers, stylists, and consumers.
This AI-driven approach blends fashion and technology, helping users create well-coordinated outfits that align with
modern styling trends. Issues like complexities in text-to-image development process and accessibility of data are
encountered. In coming time improvements includes making bigger dataset to get more colors patterns and textures, good
tuning models for fashion targeting tasks, developing automated testing frameworks for accessory matching. The model
holds energy for development into e-commerce platforms, individual styling applications and fashion suggestion systems,
giving personalized and dynamic fashion solutions.