基于专利动态指标的新兴技术预测建模方法--以癌症药物领域为例

Formulation of a Predictive Modeling Approach for Emerging Technologies: Based on Patent Dynamic Indicators in the Cancer Drug Area

新兴技术是技术创新管理、科技政策制定和技术竞争情报研究领域关注的重点问题,随着大数据挖掘技术的兴起,针对海量专利数据的自动化新兴技术识别与预测,逐渐成为当前的研究热点。本研究的核心目的是识别未来哪些癌症药物专利最有可能成为新兴技术。本研究的主要贡献是:首先,是从前向视角(exante)进行预测而不是采用回溯视角;其次,针对癌症药物领域新兴技术预测面临的数据不均衡、早期预测需求提供了一套完整的分析框架。最终,通过研究证实了利用监督学习方法构建的包含动态专利指标的模型对新兴技术具有较好的预测效果,尤其是,动态指标“前向专利家族”和“前向引文”对增强模型的效果具有显著作用。

The automated identification and prediction of emerging technologies based on massive patent data have be‐come a hot topic of research with the increasing use of big data mining technology.The purpose of this study is to identify which cancer drug patents are most likely to become emerging technologies in the future.This study has two main contri‐butions:first,the predictions are based on an ex-ante perspective instead of a retrospective perspective;second,it provides a complete analytical framework to address the issues of unbalanced data and early-stage forecast requirements for emerg‐ing technology predictions in the cancer drug field.Finally,it confirms that a model constructed on the basis of supervised learning methods that include dynamic patent metrics has good predictive power for emerging technologies.The dynamic indicators“forward patent family”and“forward citation”can effectively enhance the model’s performance.