Enhancing food freshness detection through IoT and machine learning | |
| Author | Singarapu, Vidath |
| Call Number | AIT RSPR no.IOT-24-01 |
| Subject(s) | Internet of Things Machine learning |
| Note | A research study submitted in partial fulfillment of the requirements for the degree of Master of Engineering in Internet of Things (IoT) Systems Engineering |
| Publisher | Asian Institute of Technology |
| Series Statement | |
| Abstract | In today’s world, food spoilage poses a significant challenge as it can have adverse ef fects on consumer health. Our project aims to address this issue by utilizing appropriate sensors to detect spoiled food and monitoring the gases emitted by specific food items. Through the use of a microcontroller integrated with Internet of Things (IoT) capabil ities, timely alerts are issued to prompt necessary actions. This technology holds vast potential for application in the food industry, particularly in cases where manual food detection methods are currently employed. Furthermore, we intend to incorporate ma chine learning techniques into our model to estimate the probability of food spoilage and the duration of spoilage for items purchased from different vendors. Additionally, we anticipate that our project will yield significant insights into the efficacy of combining sensor-based detection and machine learning for addressing food spoilage concerns. By enhancing our understanding of the relationships between gas emissions, spoilage indicators, and vendor sources, we can provide valuable guidance for future efforts in preventing food spoilage and promoting food safety.This a comprehensive evaluation of our food freshness detection system, highlighting the remarkable performance of the Random Forest Classifier and Support Vector Clas sifier models. With an exceptional validation accuracy of 99.58%, the Random Forest Classifier emerges as the clear top performer, demonstrating unmatched robustness and precision in classifying instances. Despite the commendable 88.33% accuracy achieved by the Support Vector Classifier, the substantial superiority of the Random Forest Clas sifier establishes it as the preferred choice for our application.Beyond statistical metrics, this conclusion emphasizes the Random Forest Classifier’s excellence not only in accuracy but also in ensuring a crucial level of dependability for the system’s success. Navigating the complexities of machine learning models, this definitive finding becomes a cornerstone for effective model selection, ensuring the un wavering reliability and precision of our food freshness detection system. The abstract encapsulates the critical role of robust model selection in seamlessly integrating ad vanced technologies for real-world applications, providing a concise overview of the pivotal outcomes derived from our comparative analysis. |
| Year | 2024 |
| Type | Research Study Project Report (RSPR) |
| School | School of Engineering and Technology |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Internet of Things (IoT) Systems Engineering |
| Chairperson(s) | Attaphongse Taparugssanagorn |
| Examination Committee(s) | Chaklam Silpasuwanchai;Chantri Polprasert |
| Scholarship Donor(s) | AIT Scholarships |
| Degree | Research report (M. Eng.) - Asian Institute of Technology, 2024 |