In real-world clinical settings, the growing number of patients and the shortage of experienced ophthalmologists make early and accurate diagnosis of retinal diseases increasingly challenging. Cataracts, diabetic retinopathy, and glaucoma are some of the most common causes of lifelong blindness around the world. This is why there is a need for automated diagnostic systems that can accurately diagnose and interpret clinical data. The major goal of this work is to find out if a deep learning architecture based on EfficientNetB3 and Explainable Artificial Intelligence (XAI) can accurately classify multiple types of retinal diseases while still being clear to doctors. The proposed system categorizes retinal fundus images into four groups: cataract, diabetic retinopathy, glaucoma, and normal. The dataset consisted of a balanced and publicly accessible collection of 4,217 retinal fundus pictures, processed using standard preprocessing techniques to enhance their generalizability. We chose EfficientNetB3 as the main architecture since it is better at extracting features, and we compared it to the standard convolutional neural network baselines to show how useful it is. The suggested model was 97% accurate in classifying better than Residual Network 50 (ResNet50) is 91% and Visual Geometry Group 16 (VGG 16) is 87%. The high precision, recall, and F1-scores (0.94 – 1.00), the Cohen’s kappa of 0.95, and the low logarithmic loss of 0.10 all point to reliable predictions. The receiver operating characteristic analysis yielded an AUC of 1.00 across all illness categories. Gradient-weighted Class Activation Mapping (Grad-CAM) was utilized to address the interpretability deficit in deep learning-based medical systems and to pinpoint clinically significant retinal regions that influence model predictions. The results indicate that employing XAI alongside EfficientNetB3 enhances both diagnostic precision and interpretability, hence validating its suitability as a transparent decision-support system for the automated screening of retinal disorders. 

In recent years, the rapid advancement of machine learning (ML) has surpassed many expectations, and its application in the healthcare sector has emerged as one of the most fascinating areas of exploration. This thesis looks into whether machine learning can increase the precision and efficacy of breast cancer diagnosis. With the help of nine classification algorithms including Random Forest, XGBoost and MLP Classifier the given work intends to propose a reliable automatic solution for malignant and benign classification of breast tumor. The main idea of the project is the development of the Web based tool that would allow doctors and other medical practitioners to make quick decisions The MLP Classifier was found to be the optimal solution after its efficiency was evaluated based on the accuracy rate, and such parameters as precision rate, recall rate, and F1-score. This leads to development of a user friendly app; even those that would not originally consider themselves technical can easily operate the application. Apart from addressing the matter of high accuracy of diagnostics, the system shows the possibility of minimizing the rates of human factors and optimizing clinical decision. Seeking for that day when technology and human opinion will complement each other in the delivery of healthcare, our study neither only contributes to the growing literature on applying artificial intelligence in healthcare but also evolves the blueprint to integrate ML models in everyday practice.

The livestock sector is an essential component of the global economy. Farmers are losing interest in this profession as animals suffer from a variety of bad health conditions, unpredictable fatal illnesses. The temperature and humidity of the farm have a stronger impact on the health of the cattle. Monitoring the health of dairy cattle and the environmental state of farms can assist to tackle these concerns. In this research, we describe a system that allows farmers to monitor livestock health metrics including body temperature as well as environmental data like temperature, humidity, light levels, quality of air, and dampness. The device will automatically maintain optimum environmental conditions for the livestock in addition to monitoring environmental parameters. Our suggested method would assure adequate water supply to the cattle in order to increase milk production supply. Three ESP32 microcontrollers are utilized as clients in this system to sense different health and environmental aspects, and an ESP32 microcontroller is used as a server to wirelessly link the three ESP32 clients. Through a web application, health and environmental factors may be accessed on the internet. It may also be accessed remotely on a mobile phone. This revolution in advanced technological farm automation will help to improve productivity by reducing the need for human intervention. Finally, the proposed solution would assist in boosting productivity while saving the farmer's time and effort.

Optimizing energy management for household appliances is essential for maximizing domestic energy utilization and enabling preventive maintenance. Recent studies indicate that traditional forecasting approaches frequently lack the necessary accuracy and real-time learning capabilities required for effective management of household energy. This study demonstrates the implementation of a comprehensive strategy that integrates Internet of Things (IoT) data, machine learning (ML), and explainable artificial intelligence (XAI) to improve the accuracy and interpretability of predicting energy usage in residential buildings. Our research focuses on the rising issues faced by IoT-based smart systems, partic- ularly the deficiencies in the performance of current solutions. Therefore, as compared to the other 17 models that were examined, polynomial regression demonstrated outstanding performance. Our solution utilizes a non-intrusive sensor to collect data without disrupting its operation. Real-time data collecting is achieved through a Flask-based web page with Ngrok for external access.The efficacy of the proposed system was assessed using many metrics, yielding highly satisfac- tory results: the root mean square error (RMSE) was 0.03, the mean absolute error (MAE) was 0.02, the mean absolute percentage error (MAPE) was 0.04, and the coefficient of determination (R²) was 0.9989. However, modern cutting-edge methods still face considerable hurdles when it comes to interpretability. In order to tackle these problems, we include XAI techniques such as SHAP and LIME. Explainable Artificial Intelligence (XAI) improves the interpretability of the model by elucidating the impact of various variables on energy consumption forecasts. Not only does this increase the effectiveness of the model, but it also promotes comprehension of the data and enables them to identify the elements that influence home energy usage.