Text mining for technology monitoring
Thorsten Teichert, Marc-André Mittermayer · 2003
A considerable part of scientific and technological knowledge is coded in writing. In this context, automated text categorization can be regarded as a promising tool particularly for patent data analysis. In a real-life example, we show that automated text categorization can closely resemble the time-consuming categorisation job of an expert. By comparing different algorithms we reveal systematic differences in their results and show potential for further improvement.