New Hybrid Algorithm Implementation on spread Wireless Sensor to determine the point of fire in the building

This research focused on developing accuracy to determine the fire point of a building using Wireless Sensor Network (WSN) which had been spread unevenly and in motion. At present, the constraints faced by WSN that are distributed unevenly are the weak level of accuracy of the data obtained, which can be caused by the weakness of the WSN itself, namely resource limitations in the form of bandwidth, power, delay, latency, and computing. Therefore, an intelligent algorithm that functions to process data based on the unevenly distributed WSN was needed to get better accuracy in building fire determination information. This study used 2 types of algorithms, namely K-Nearest Neighbor(k-NN) for the grouping of positions from the scattered WSN, and the Decision Tree algorithm to determine the outcome of a building fire decision based on each of the grouping results. From testing of 20 WSNs that implemented the 2 algorithms mentioned earlier, the results obtained were in the form of an accuracy rate of 70% and an average error of 30%.