1、TECHNOLOGY SCAN:HUMAN&MACHINE COLLABORATION WITH INPUT FROM METHODOLOGY 1.Dataset used for the report The patent dataset was retrieved on 17 May 2019 and comprises worldwide patent applications relating to human and machine collaboration technologies published in 2009-2018.Relevant business informat

2、ion,market data,and national policies that are available from commercial databases or on the web are also used to support the findings of the report.2.Counting the number of inventions This report counts the number of inventions by the number of unique patent families.Counting individual patent appl

3、ications will result in double counting as each patent family may contain several patent publications if the applicant files the same invention for patent protection in multiple destinations.As a patent family is a group of patent applications relating to the same invention,analyses based on countin

4、g one invention per unique patent family can reflect innovation activity more accurately.3.Formulation of search strings To ensure optimal recall and accuracy of the data sets retrieved,the search strings used in this study were formulated by incorporating keywords(and their variants),as well as rel

5、evant patent classification codes and indexes,e.g.International Patent Classification(IPC)and Cooperative Patent Classification(CPC).4.Grouping of technology domains Grouping of individual patent documents into the respective technology domains was carried out based on patent classifications codes,t

6、ext-mining and semantic analysis of the patent specifications in particular claims,titles,abstracts,as well as a manual review of the individual patent applications.5.Growth rate calculation Annual growth rate refers to the average annual growth and was derived by using the best-fit exponential line