Recommender system using decision tree and neural network for disaster risk reduction management services

ABSTRACT: At the occurrence of natural disasters alongside other environmental crises, the Disaster Risk Reduction Management Services (DRRMS) of the Department of Education Central Office upholds the initiative of responding to these events within schools of their responsibility. In collaboration with the DRRMS, the affected institutions most generate a Rapid Assessment of Damage Report (RADaR) to relay a detailed account to the central office regarding the overall condition of each school and its population. However, the DRRMS with remarkable years of efficiency may still experience difficulty in analyzing RADaR files. A particular problem raised in this study is which academic institutions bear the most damage and must be prioritized for response/recovery plans. On top of that, what specific measures must be implemented to ensure the school’s recovery from the disaster? Hence, this study aims to utilize data mining through a Recommender System empowered by Decision Tree, Neural Networks, and Content-Based Filtering. The developed product accepts a standardized RADaR file and from these will determine the severity via Decision Tree, provide a detailed assessment through Neural Network, and generate recommendations using the Content-Based Filtering. It also optimizes as SMS notification feature in which the DRRMS can easily relay recommendations to the affected schools. To sum it all up, this Recommender System cases data processing amidst extensive loads of RADaR files and has proven to further improve accuracy, reliability, and productivity.

Filipiniana