PIR Sensor-Based Intelligent Home Security System Design Analysis Using Machine Learning Methods

Home security is now a major concern, not only in big cities, but also in suburban areas, especially for people who carry out many activities outside the home. Strategies are needed to improve and secure security and surveillance systems around the home that can be monitored remotely. Smart home security with motion detectors is her one of the solutions to this problem. In this study, we design and analyze a motion detector for integrated home security using four passive infrared sensors. When the passive infrared sensor module detects motion, the detection results will provide a motion detection notification, which will be sent to the homeowner’s smartphone via the Blynk application. The dataset was split 70:30. Dataset test was performed using the bagging method to determine the data matrix. The test results revealed the highest values for the bagging method in Confusion Matrix calculations before and after tuning. The bagging method has five classifiers: Random Forest, Extra Tree, K-Nearest Neighbors (KNN), Support Vector Classifier (SVC), and Ridge, and requires more computation time compared to using fewer classifiers. The confusion matrix values for the test data after tuning show 95.7% for accuracy, precision, recall, and F1 score. The results show that the proposed prototype can be used as a home security model.