CATMInE Computer Assisted Trade-Mark Infringement Evaluation

In forensic Machine Learning, information systems are not easy to use due to many constraints. Knowledge generally comes from laws and their use is represented by decisions. The whole data requires several parses in order to obtain a structured database on which decision tools can be applied. This study presents the use of machine learning tools to assist the lawyer in nominative trade-mark infringement evaluation. The domain formalism and database are obtained from the JURINPI database which regroups thousands of cases in the trade-mark domain. The use of a neural network and of Sammon’s non-linear mapping provides a development of datas with designed decision tools for trade-mark infringement evaluation : CATMInE (Computer Assisted Trade-Mark Infringement Evaluation). This web-oriented application requires little technical knowledge on machine learning and displays a readable and concise information.