A Healthcare Decision Support System assists healthcare practitioners by systematically analyzing vital health information and making meaningful decisions for quality patient care. Clinical and non-clinical healthcare decisions are made in various ways, with the most critical areas being disease diagnosis (clinical), treatment management (clinical) and resource management (non-clinical). However, decision-making is very complicated in these areas because of their sensitivity and diversity. But it’s very significant towards the stakeholders as physicians must diagnose the disease and make rapid treatment decisions for patients, while hospitals must provide appropriate treatment with limited resources as well as earning maximum profit. Such kind of healthcare decision support system becomes more significant in least developed or developing countries, where healthcare resources are in limited supply and also their management system is mostly manual. In such circumstances, the Intelligent Healthcare Decision Support System (IHDSS) can be an effective technique, allowing the utilization of computational intelligence to make decisions for a range of healthcare concerns employing relevant health data.

Cognitive Computing a set of contextual, interactive and adaptive systems that adds a human touch to user experiences online. Cognitive computing in e-commerce is more instinctive and hence successful in collecting rich personal data. Cognitive technology not only empowers consumers but also helps them to re-discover themselves. E-commerce has increasingly become an important part of business strategy in the emerging global economy due to its 24/7 availability, rapid access, wide selection, and international reach for consumers. E-commerce management needs artificial autonomous systems to solve the cognitive problems associated with human intelligence, deal with customer data, forecast customer behavior, and process large volumes of data. In every aspect of your e-commerce operation, artificial Intelligence and machine learning strategies offer business benefits, especially when it comes to predicting customer churn and retention. This research project aims to develop an AI driven integrated framework for E-commerce management by utilizing some state of the art cognitive computing techniques.

A smart tolling system represents a revolutionary approach to toll fee collection, utilizing cutting-edge technologies such as computer vision and cloud-based microservices. Manual toll fee collection at bridges in Bangladesh leads to traffic congestion, resource intensive operations, and errors. This study presents a pioneering framework that leverages computer vision and cloud-based microservices to transform toll fee collection. It marks Bangladesh’s first computer vision and cloud-based microservices implementation for automated toll fee collection. Our system utilizes advanced computer vision technology, primarily focusing on the YOLO The tiny v4 model enables efficient and precise vehicle detection from a remarkable one-meter distance, eliminating the need for vehicles to stop at toll gates. Furthermore, our system integrates a separate Optical Character Recognition (OCR) the engine that meticulously processes the top captured license plate image based on confidence scores for accurately recognizing Bengali characters. The system categorizes vehicles using information from Bangladesh Road Transport Authority (BRTA)-approved license plates and automates toll fee collection through an integrated payment server. Successful payments trigger automatic toll gate openings, streamlining operations. Our system achieves impressive accuracy rates, with 91% accuracy in license plate recognition against ground truth data.

Human activity recognition (HAR) is becoming extensively popular due to its abundant and far-reaching applications in smart homes, monitoring, and surveillance. HAR offers valuable insights into a person's physical functioning and behavior, which can be automatically monitored to provide individualized assistance. With the advancement of Wi-Fi technologies, individuals are now surrounded by devices capable of sensing and communication, which makes activity identification significantly more efficient than image/video-based and wearable sensors-based approaches.