Bar-Ilan University, Ramat Gan, Israel.
PLoS One. 2022 Feb 11;17(2):e0263891. doi: 10.1371/journal.pone.0263891. eCollection 2022.
Crowdfunding platforms allow entrepreneurs to publish projects and raise funds for realizing them. Hence, the question of what influences projects' fundraising success is very important. Previous studies examined various factors such as project goals and project duration that may influence the outcomes of fundraising campaigns. We present a novel model for predicting the success of crowdfunding projects in meeting their funding goals. Our model focuses on semantic features only, whose performance is comparable to that of previous models. In an additional model we developed, we examine both project metadata and project semantics, delivering a comprehensive study of factors influencing crowdfunding success. Further, we analyze a large dataset of crowdfunding project data, larger than reported in the art. Finally, we show that when combining semantics and metadata, we arrive at F1 score accuracy of 96.2%. We compare our model's accuracy to the accuracy of previous research models by applying their methods on our dataset, and demonstrate higher accuracy of our model. In addition to our scientific contribution, we provide practical recommendations that may increase project funding success chances.
众筹平台允许创业者发布项目并为实现项目筹集资金。因此，影响项目筹款成功的因素是非常重要的。以前的研究考察了各种因素，如项目目标和项目持续时间，这些因素可能会影响筹款活动的结果。我们提出了一种新的模型来预测众筹项目实现资金目标的成功率。我们的模型只关注语义特征，其性能与以前的模型相当。在我们开发的另一个模型中，我们同时考察了项目元数据和项目语义，全面研究了影响众筹成功的因素。此外，我们分析了一个比以前报道的更大的众筹项目数据集。最后，我们表明，当结合语义和元数据时，我们的 F1 分数准确率达到了 96.2%。我们通过在我们的数据集上应用他们的方法来比较我们的模型的准确性和以前的研究模型的准确性，并展示了我们的模型的更高的准确性。除了我们的科学贡献，我们还提供了一些实用的建议，可以增加项目筹款成功的机会。
