Published In

2019 Portland International Conference on Management of Engineering and Technology (PICMET)

Document Type


Publication Date



Technology -- Management, Crowd funding -- Decision making, Keyword searching -- Analysis, Natural language processing,


In the technology management field, informal source of information such as social media has been used for technology mining, leader user detection, and etc. However, usually unstructured nature of the informal information introduces some challenges. This study is part of an effort to automate key product features extraction from crowdfunding campaign's textual information, in order to analyze the effect of them on the decision-making process of crowdfunding backers. This paper intends to evaluate the performance of UDPipe R package and six keyword extraction techniques on candidate features selection from textual data of 3D printer campaigns on Kickstarter. In the end, it will be discussed which technique captures the 3D printer features better.


This is the publisher's final PDF. Copyright 2019 by PICMET. Paper delivered at the 2019 Proceedings of PICMET '19: Technology Management in the World of Intelligent Systems



Persistent Identifier

Included in

Social Media Commons