Open AccessBridge Construction Cost Prediction using Multiple Linear Regression
Author(s) -
Krishna Garkal,
N.B. Chaphalkar,
Sayali Sandbhor
Publication year - 2019
Publication title -
international journal of innovative technology and exploring engineering
Language(s) - English
Resource type - Journals
ISSN - 2278-3075
DOI - 10.35940/ijitee.i8916.078919
Subject(s) - linear regression , mean absolute percentage error , statistics , variables , regression analysis , mathematics , pearson product moment correlation coefficient , regression , bridge (graph theory) , variable (mathematics) , mean squared error , medicine , mathematical analysis
Cost of construction of bridges is predicted using multiple linear regression model, based on data of bridges from Maharashtra state in India. Cost per unit area is taken as an appropriate dependent variable. Using both conventional and double log regression techniques, models for cost/m2 and log of cost/m2 are developed. Total 6 independent variables, which include both qualitative and quantitative variables, are used to develop the model. Height of bridge, average span length and depth of foundation are used as quantitative variables. Zone of construction, deck type and foundation type are used as qualitative variables in developing model. Strength of these independent variables with dependent variable is found out using pearson’s correlation method. Model is then verified using Leave One Out Cross Validation (LOOCV) technique. The most suited regression model obtained from the data experiment is double log regression with R2 of 0.850 and a Mean Absolute Percentage Error (MAPE) of 17.74%, as compared to 25% MAPE observed in past for studies related to traditional cost prediction.