Cost Estimation Using Function Point for an AIoT-Based Infusion Drop Detection Prototype Using Random Forest and XGBoost

This study aims to develop an Artificial Intelligence of Things (AIoT) prototype capable of accurately and real-time predicting the drip rate (TPM) of intravenous fluid administered to patients. The research focuses on three main aspects: the design, implementation, and testing of an Internet of Things (IoT)-based prototype that integrates sensors with a microcontroller to detect fluid drips; the development of machine learning models using the Random Forest and XGBoost algorithms to classify drip status based on sensor data; and the estimation of software development costs using the Function Point (FP) IFPUG method. The data and system requirements were obtained through a literature review of previous studies on IoT-based infusion prototypes and online health references to determine the drip calculation formula used for dataset construction. To achieve these objectives, the study combines three development approaches. The Throw Away Prototyping method was applied in designing and constructing the AIoT hardware, while a machine learning pipeline was used to train and evaluate the TPM prediction model. Furthermore, the Function Point IFPUG method was employed to estimate the software development costs by considering the accuracy level of the developed model. The findings of this study are expected to contribute to the advancement of AIoT-based infusion monitoring systems that are adaptive and applicable within healthcare service environments.