CBR基盤 回歸分析技法을 利用한 道路事業 初期段階 工事費 豫測모델 開發

Abstract

The prediction of optimal construction expenses in the initial stage of public construction projects is an essential part of successful project execution and budget management, and first of all, it is very important in decision-making of the initial stage of a project. However, the available information such as design drawings and construction specification, etc. in the construction planning stage is informations in an uncertain state, so its content is being limited. In relation with this, construction expenses suggested after being estimated in a state that the adequacy is not secured are connected unreasonable design and construction so that there is worry that will change the project nature. The prediction of current construction expenses is following a method being suggested by the Ministry of Land, Transport and Maritime Affairs and Korea Development Institute, and this is being progressed in a method that calculates through utilization of construction expenses per unit length of simple linear formula based on previously constructed project materials. Considering this, this research has referred previously executed project materials, but suggested an algorithm for raising prediction efficiency of construction expenses by being deviated from simple statistical methods, and performed development and verification of a prediction model of construction expenses for enhancing adequacy. This research has carried out utilization of acquirable information as a means in the planning stage by collecting materials of construction expenses from 2003 to 2010 in terms of road projects previously performed or being enforced inside the Jeju Special Self-governing Island. The factors of road project's construction expenses suggested by this research were used for extraction of similar cases through correlation analysis by deducing preceding research literature and expert counselling, and utilized it as a prediction model of construction expenses based on case-based reasoning by suggesting influence(weight) of factors through regression analysis. In addition, by comparing it with a prediction model of construction expenses suggested by the Ministry of Land, Transport and Maritime Affairs and Korea Development Institute which corresponds to the current estimation standard, this research aims to suggest a more accurate prediction model of construction expenses than the current initial stage's prediction model of construction expenses by verifying reliability of a prediction model of the initial stage's road project construction expenses suggested from this research.