Comparative Analysis for Black Spot Identification Methods: Combined Methods in Producing Heat Map for Black Spots within Macarthur Highway, Jose Abad Santos Avenue, and Lazatin Boulevard in the City of San Fernando, Pampanga, Philippines

One of the leading causes of mortality in the world is road accidents. As a matter of fact, it is the 8th leading cause of death worldwide.High number of road accidents in a specific zone of a road or stretch is commonly term as black spot. In the Philippines, this is also an issue,specifically in its cities. The study area was the City of San Fernando, Pampanga of Philippines. In determining the black spot, an analysis wasdue, but there were a lot of methods used. The researchers conducted a comparative analysis between the models and created heat maps basedon the objectives of the researchers. The researchers only selected a few of the models that were possible to perform within the given timeframe and data availability. The classification of models was divided into two: Probability model and Severity model. Under Probabilitymodel, Poisson Regression and Grey-Verhulst were utilized. Under Severity model, Rate Quality Control and Empirical Bayesian wereutilized. These were applied on some portions of the three stretches of roads of City of San Fernando, Pampanga, namely MacArthurHighway, Jose Abad Santos Avenue (JASA), and Lazatin Boulevard. To test the accuracy of probability model, Relative Error and Root MeanSquare Error (RMSE) were used. The probability model with lower error was used to combine with severity model allowing the researchers tocreate two heat maps. It was concluded that in consideration of the data collected, Grey-Verhulst is accurate than Poisson Regression showingan error of 12.93% and 7.49, and 22.25% and 8.26, respectively, under Relative Error and RMSE, respectively. This showed that GreyVerhulst showed a relatively low of error with a difference of 9.32% and 0.67 to Relative Error and RMSE, respectively of PoissonRegression. As a result, Grey-Verhulst was combined to Rate Quality Control and Empirical Bayesian, to accomplish and create the heatmaps.