Computational Approach to Track Cold Chain Product in Web Application

Abstract

In cold chain logistics sector, tracking the temperature of perishable is major challenge with the quality and safety of the products. Food products, pharmaceuticals, specialty chemicals are stored in cold storage, which are categorised as perishable and non-perishable items. Between manufacturing and marketing, cold storage is the most efficient location for storing perishable items in bulk, particularly fruits and vegetables. A decision tree is a supervised learning algorithm in machine learning that is perfect for classification problems, as it is able to order classes on a precise level. This creates categories within categories, allowing for classification with limited human supervision. The algorithm is used in this work to label and categorize the unlabelled data that is obtained from the client. The labelled data is then analysed to determine certain factors like freshness, loss quantity, quantity in fresh and other characteristics. Hence this work focuses on computational techniques like decision tree and support vector machine algorithm, to identify the products based on different set of features. User can receive quick information about the state of the items that are kept in cold storage based on these determined factors. Food products and other perishable items require the most supervision since their degradation occurs more quickly. Web based support ensures the quality and efficient quantity in fresh which is useful for clients knowing the information about their products at all times. Compare to conventional approach, decision tree-based classification is able to track the product with highest accuracy.