Introducing DynaPTI-constructing a dynamic patent technology indicator using text mining and machine learning.

Area of Science:

• Innovation Studies
• Intellectual Property Management
• Technology Indicators

Background:

• Patent data is crucial for research and corporate intelligence.
• Current patent-based indicators often overlook firm-level dynamics in technological quality and activity.
• This omission leads to incomplete and potentially biased assessments of firm-level innovation.

Purpose of the Study:

• To develop a novel patent-based technology indicator, DynaPTI (Dynamic Patent-based Technology Indicator).
• To address the limitations of existing indicators by incorporating firm-level dynamics and textual data.
• To provide more precise and up-to-date assessments of firm-level innovation activities.

Main Methods:

• Developed a dynamic, index-based comparison framework for firms.
• Integrated machine learning techniques to analyze textual information from patent documents.
• Applied the DynaPTI framework to companies in the wind energy sector for empirical validation.

Main Results:

• The DynaPTI framework provides more accurate and timely assessments of firm-level innovation.
• Empirical results demonstrate the indicator's ability to identify emerging innovation overperformers.
• Findings suggest DynaPTI offers valuable insights complementary to existing approaches.

Conclusions:

• DynaPTI enhances the analysis of patent data by incorporating dynamic firm-level factors.
• The indicator offers a more comprehensive view of technological activity and quality.
• This approach is particularly useful for identifying nascent leaders in specific technological fields.