Modeling of Incident Severity Classification on Fuel Tanker Truck Using k-NN

The frequent occurrence of fuel tanker truck accidents in distributing fuel oil needs attention and must be anticipated early. Some factors that cause fuel tanker truck accidents are human error, road problems, environmental conditions, and carrier and tank design. Evaluation and anticipation of accidents are necessary for the Modeling of Incident Severity Classification to create policies to prevent accidents on fuel tanker trucks. Modeling of Incident Severity classification using k-nearest neighbors (k-NN) by classifying minor, insignificant, significant, catastrophic, and moderate incidents. The results of this modeling obtained the values of k-3 and k-5 with the same accuracy value of 92.31%. At the value k-3, it produces a precision minor of 100<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup>, insignificant 94.12%, and significant 100, while recall minor is 85.71%, insignificant 100%, and significant 66.67%. At the value of k-7, it produces a precision minor of 100<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup>, insignificant 88.89%, and significant 100%, while recall minor is 71.43%, insignificant 100%, and significant 100%. The results of this modeling can be used as a reference for policymaking to reduce and prevent accidents on fuel tanker trucks.