Date

5-2019

Degree

Bachelor of Science in Civil Engineering

College

College of Engineering and Agro-Industrial Technology (CEAT)

Adviser/Committee Chair

Sundo, Marloe B.

Abstract

This study presents the development of a safety performance function for arterial intersection in Metro Manila, Philippines, using Random Decision Forest which can be used to determine accident rates at an intersection based on its characteristics. A heat map was generated which shows that the intersection along Don Marcos Avenue in Quezon City has the most number of recorded accidents from 2013-2o15. From the Random Decision Forest, the significant variables affecting accident counts are number of lanes in major and minor roads, presence of separate pass, presence of pedestrian lanes, and type of land use. However, upon performing the negative binomial regression analysis, the number of lanes in minor roads was found to be insignificant, thus, its exclusion from the equation. The safety performance function?s sensitivity based on train data is 68.3% while its specificity is 69.1%. Significant difference was observed upon the incorporation of test data into the model due to overdispersion. Nevertheless, the Akaike Information Criterion shows that the safety performance function generated is the most suitable model for the arterial intersections in Metro Manila, and the sensitivity and specificity values obtained are within acceptable range, thus, the safety performance function developed is acceptable.

Language

English

LC Subject

Roads--Interchanges and intersections. Traffic safety. Traffic accidents.

Location

UPLB College of Engineering and Agro-Industrial Technology (CEAT)

Call Number

LG 993.5 2019 E63 /L37

Document Type

Thesis

Download

Share

COinS