Experience mining based on text analytics and case-based reasoning to support crowdfunding design

Crowdfunding has emerged as an important financing channel. Startups can raise funds through crowdfunding platforms. Owing to the limited experience and uncertainty of market demand, designing crowdfunding projects to attract online investors is often confusing for start-ups. This study proposes an integrated model of text analytics and case-based reasoning that assists in experience mining when designing crowdfunding projects. Case representation and retrieval are the two core components of the model. In addition to the literature review, text analytics methods, such as latent Dirichlet allocation and Doc2Vec model based on neural network, are also used to extract features to represent crowdfunding cases. For features with three value formats, namely numerical, categorical, and text vector features, three local similarity measures are adopted for case retrieval. Finally, the effectiveness of the model was verified using project data on the Kickstarter platform. The results show that the proposed model provides fundraisers with valuable experience in improving their design schemes and obtaining better financing performance.

Automated summary

Crowdfunding has become an important financing channel for startups. This study proposes an integrated model of text analytics and case-based reasoning to assist in experience mining when designing crowdfunding projects. The model utilizes text analytics methods to extract features and adopts local similarity measures for case retrieval. The effectiveness of the model is verified using project data on the Kickstarter platform, demonstrating its value in improving design schemes and financing performance for fundraisers.