| Summary: | 碩士 === 國立高雄應用科技大學 === 土木工程與防災科技研究所 === 98 === As the building construction projects become larger and more complex, researchers and the industry practitioners both recognize the importance of early planning and its potential impact to final project outcome. Nevertheless, the preproject planning practice varies significantly throughout the industry. This research incorporates Project Definition Rating Index (PDRI) as part of the survey questionnaire to investigate the current practice of preproject planning for the building construction in Taiwan. Information from a total of 92 building projects is collected to study the relationship between the level of preproject planning and project performances (cost and schedule). With the PDRI scores as the independent variable and project performances (cost and schedule separately) as the dependent variable, statistical methods (Logistic Regression) and Artificial Neural Network (Bootstrap Aggregating and AdaBoost) methods are used to develop prediction models. From the analysis, it shows that AdaBoost method yields the best prediction results for both the cost and schedule models (84% accuracy for cost performance and 80% accuracy for schedule performance predictions). The research results show that the level of preproject planning has positive on project outcomes. That is, projects with better preproject planning are more likely to have better project performance (cost and schedule). In addition, the PDRI scores can be used to predict project performances using statistical and ANN models. The research results provide valuable information for future preproject planning researches in Taiwan. The industry practitioners can utilize the results to improve their preproject planning process and hence enhance their chance of project success.
|
|---|
