[Column] Introduction to TM-RoBo's Features: Title Search - 2

Create a learning model by comprehensively using machine learning to judge the classification of titles related to the Patent Office and courts!

In the previous installment, we introduced the background issues of existing trademark database services that led to the creation of TM-RoBo's phonetic search. This time, we will explain "the mechanism of phonetic search." As mentioned previously, the results of phonetic searches in existing trademark database services either randomly list trademarks that hit the search or only arrange them according to generally known formal rules (such as a one-sound difference indicating high similarity), so they are not ordered in a correct similarity sequence. As a result, it was necessary for researchers to carefully make similarity judgments for all search results themselves. *For more detailed content of the column, please refer to the related links. If you have any questions, feel free to contact us.*

Database