Predicting Regional Development Competitiveness Index Using Naive Bayes: Basis for Recommender System

Abstract

Competitive index is the diagnostic tool in assessing competitiveness of the city or municipality to determine problematic factors for improvement. Naive Bayes algorithm is a useful method in predicting development competitive index basis for recommendation in the development plan. The predicted status can provide vision for development plan, business investment, policy making, and resiliency to calamities. Specifically, it addressed the following objectives: (1) indicated all the indicators in the competitiveness index for the different cities and municipalities; (2) computed the relative efficiency among cities and municipalities using Data Envelopment Analysis (DEA); (3) ranked the relative efficiency to determine competitiveness of a city or municipality in Region 1; and (4) utilized the Bayes theorem to determine the probability of competitiveness of all the city or municipality in Region 1 basis for recommendation system. The study classifies Local Government Units (LGU’s) competitiveness index based on the four pillars from cities and municipalities competitive index survey tool. The four pillars focuses on government efficiency, economic dynamism, infrastructure and resiliency. In the process of development, it involves applied and developmental research designs, Cross-Industry Standard Process for Data Mining (CRISP-DM) Data Envelopment Analysis and Navie Bayes algorithm. The CRISP-DM is the method used in preparing and processing of Data. While Data Envelopment Analysis was used to determine the relative efficiency, target, slack and slack percentage. Using Waikato Environment for Knowledge Analysis (WEKA) as a tool for prediction applying the Navie Bayes algorithm, the accuracy result is 90.64 percent. Utilizing the Navie Bayes algorithm determined the probability of competitiveness of all LGU’s in Region 1 for recommendations in the development plan.