近年來，印尼面臨在基樁工程生產力減損之問題，在增加基樁工程生產力之前，辨識出其影響因子及其影響程度為首要任務。本研究目標為辨識影響基樁工程之巨量因子，藉由支援向量回歸機精準預測其生產力之減損，並從相似之案例得知其可能耗損之數量。 文獻回顧對於支援向量回歸機指出5項巨量因子(勞工、管理、環境、材料及設備)及8種投入項目(土壤情況、基樁種類、基樁材料、專案大小、專案所在地、基樁深度、基樁數量及設備數量)，並由以上所提出之因子及項目找出在印尼爪哇島5個主要地區共110項基樁工程專案。支援向量回歸機經10次交叉驗證後得到87.2% 的精確度，並由以往相似案例可得知生產力的減損大約占總生產力之18.55%。調查結果將使從事此作業者更加注意減損問題以增加整體生產力。;Pile construction productivity loss in Indonesia had been occurred for years. Before improving pile construction productivity, impact factors and how much potential loss are urgent to identified. The research objectives are to identify the macro factors that influence pile construction, to develop a SVR model that precisely predicts productivity loss, and to provide potential loss quantities using the most similar historical case(s). Literature review identifies 5 macro factors (labor, management, environment, material, and equipment) and 8 inputs (soil condition, pile type, pile material, project size, project location, pile depth, pile quantity, and equipment quantity) for Support Vector Regression (SVR) model, and then leads the study to collect 110 pile construction projects among 5 major areas in Java island of Indonesia. The SVR evaluated using 10-way cross validation yields an accuracy rate at 87.2%. The most likely productivity loss obtained based on the most similar historical cases is approximately 18.55% of total productivity. The findings would push the practitioners to pay attention to the loss in order to improve the overall productivity.
