Low success rates have been one of the critical issues in crowdfunding. Previous studies already indicated that the project description will affects the success of the crowdfunding project. However, there is no research to explore which factors should be included in the project description for directly affecting the success of the project. The participants’ text comments have been confirmed that they might change supporters’ decisions. In most of crowdfunding related studies, they often used questionnaires survey which may have sampling bias and require a lot of manpower and time. In addition, they merely focus on music and sports fields but have not yet discussed the movie projects. Consequently, this study attempts to present a scheme to identify the key success factors, especially for project description and user comments, using text mining and data mining approaches. In our presented scheme, feature selection methods, including Decision Trees (DT), Least Absolute Shrinkage and Selection Operator (LASSO), and Back Propagation Network (BPN) pruning method are employed to select important factors from real projects in Kickstarter and Indiegogo. Then, a support vector machine (SVM) is performed to evaluate the performances of selected candidate factor subsets. Finally, we can determine the key success factors for movie crowdfunding projects. Experimental results can give the fundraisers useful suggestions for increasing the success rate of movie crowdfunding projects.
All Science Journal Classification (ASJC) codes