TY - GEN
T1 - Analyzing risk factors affecting project cost overrun
AU - Tsunoda, Masateru
AU - Monden, Akito
AU - Matsumoto, Kenichi
AU - Hatano, Ryosuke
AU - Nakano, Toshihiko
AU - Fukuchi, Yutaka
PY - 2013
Y1 - 2013
N2 - To prevent cost overrun of software projects, it is effective to predict the project which has high risk of cost overrun in the early phase of the project. In this paper, we clarify the risk factors which affect cost overrun. The risk factors are denoted by the questions such as Are the customer's project goals clear? The risk factors can be used as independent variables of the cost overrun prediction model. In the analysis, we used 290 projects' data collected in a software development company. The dataset was stratified by the project start time and the project size to eliminate their influence, and relationships between risk factors and cost overrun were analyzed with the correlation ratio. In addition, we focused risk factors which have strong and stable relationships to cost overrun, and analyzed them using the Sharpe ratio based index. As a result, we identified some risk factors which have relatively strong and stable relationships to cost overrun. After the analysis, we experimentally predicted cost overrun projects by collaborative filtering, using the risk factors as independent variables. The result suggested that cost overrun projects can be predicted by the risk factors.
AB - To prevent cost overrun of software projects, it is effective to predict the project which has high risk of cost overrun in the early phase of the project. In this paper, we clarify the risk factors which affect cost overrun. The risk factors are denoted by the questions such as Are the customer's project goals clear? The risk factors can be used as independent variables of the cost overrun prediction model. In the analysis, we used 290 projects' data collected in a software development company. The dataset was stratified by the project start time and the project size to eliminate their influence, and relationships between risk factors and cost overrun were analyzed with the correlation ratio. In addition, we focused risk factors which have strong and stable relationships to cost overrun, and analyzed them using the Sharpe ratio based index. As a result, we identified some risk factors which have relatively strong and stable relationships to cost overrun. After the analysis, we experimentally predicted cost overrun projects by collaborative filtering, using the risk factors as independent variables. The result suggested that cost overrun projects can be predicted by the risk factors.
KW - Sharpe ratio
KW - collaborative filtering
KW - correlation ratio
KW - risk management
KW - stratification
UR - http://www.scopus.com/inward/record.url?scp=84867643976&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84867643976&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-32172-6_14
DO - 10.1007/978-3-642-32172-6_14
M3 - Conference contribution
AN - SCOPUS:84867643976
SN - 9783642321719
T3 - Studies in Computational Intelligence
SP - 171
EP - 184
BT - Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing 2012
PB - Springer Verlag
ER -