Smart Fire Safety: Analyzing Radial Basis Function Kernel in SVM for IoT-driven Smoke Detection

Abstract

This research explores the application of Support Vector Machine (SVM) with the Radial Basis Function (RBF) kernel in smoke detection using a dataset collected from Internet of Things (IoT) devices, specifically Photoelectric Smoke Detectors. With 62,630 records and 16 attributes, the study aims to address limitations in smoke detection technology that may impact system accuracy. Through RBF kernel analysis, the SVM model demonstrates the capability to recognize complex patterns related to smoke presence, achieving an accuracy rate of 96.85%. The Classification Report reveals high precision, recall, and f1-score for both "No Fire" and "Fire" detection classes. Despite encountering some false positives, particularly in specific environmental conditions, the evaluation underscores the effectiveness of the model. Recommendations include integrating the model into security systems and further exploring model development by considering environmental factors. This research provides profound insights into smoke detection and affirms its relevance in advancing superior artificial intelligence solutions. In conclusion, the SVM model with the RBF kernel proves reliable for smoke detection with broad potential applications in fire risk mitigation.

Keywords; Smoke Detection, Support Vector Machine (SVM), Radial Basis Function (RBF) Kernel, IoT Devices, Classification Report