基于专利文本数据挖掘的技术预测方法与实证研究——以纳米技术在能源领域应用为例

Technology Forecasting Based on Data Mining of Patent Texts: An Empirical Research of Nanotechnology in the Field of Energy

[目的/意义]利用文本数据挖掘技术深入挖掘专利文献主题内容,与更具有前瞻性的领域规划文本进行对比分析,发现哪些技术主题是当前饱和技术主题,哪些是未来潜在发展技术主题,以期提高技术预测的准确性,并降低预测结果的风险性。[方法/过程]在对国内外的技术预测方法进行梳理总结的基础上,利用文本数据挖掘技术挖掘专利文献技术主题,在此基础上加入时间维度,形成专利技术路线图,然后与规划文本中的规划信息进行对比分析,从而进行技术预测分析。[结果/结论]通过纳米技术在能源领域的应用进行实证研究,有效地实现了技术预测,并降低了技术预测结果的风险性。[局限]仅从专利的研究水平指标判别研究主题潜力,尚未考虑其他影响因素,指标体系还不完善;在数据源的选择上技术领域专利文本和规划文本存在不对称的情况。

[ Purpose/significance ] This paper mines the main contents of patent texts by using text data mining technologies and compares the results with the planning documents in the prospective fields. By doing this, it can discover the potential technical themes to enhance the accuracy of technology prediction and lower the risks of prediction results. [ Method/process ] The paper first reviews the technology prediction methods both in home and abroad and mines the technical themes of patent documents by using text data mining technology. Based on that, taking time dimension into consideration, the paper draws the patent technology roadmap and compares it with information in the planning documents to conduct technology forecasting. [ Result/conclusion ] The empirical research of nanotechnology in energy field shows that the method proposed in the paper effectively predicts the technology and cuts the risks of prediction results. [ Limitations ] First, the paper predicts the potential of research theme only from research status of patents without considering other influential factors, so the indicator system is not complete. Second, patent texts and planning documents in technical field could not perfectly match each other.