Smoke-Detection (Cross-Model-Validation)

A comparative analysis of traditional classification models for smoke detection.

This project consists of:

1. Smoke Detection using traditional ML classification models [Logistic Regression, K-Nearest Neighbours (KNN), Support Vector Machine (SVM), Naive-Bayes, Random Forest and eXtreme Gradient Boosting (XGBoost)].
2. A comparative analysis of model performance on the given data.

About the dataset:

This dataset consists of 60,000 readings of temperature, humidity, pressure, particulate matter, concentrations of compounds such as Hydrogen (H2), Ethanol and Carbon Dioxide (CO2), etc. taken using a set of different types of sensors from various indoor and outdoor locations as described in this hackster.io post about a real-time smoke detection system (https://www.hackster.io/stefanblattmann/real-time-smoke-detection-with-ai-based-sensor-fusion-1086e6). Data collection from several different locations along with the use of various different sensors provide us with a diverse set of features and data points to predict the presence or absence of smoke.