Blogs

Auto-Patent Categorization

Where is an industry headed? What are your competitors doing? What are the emerging trends in a technology space? These questions can be answered by analysing patent data. Technical insights derived from patent data are important for businesses to assess the strength of their technology portfolio and determine risks posed by the competitors, for academics and governments to plan the R&D investment and innovation, and for investors to perform due diligence and value companies. However, analysing patent data is an arduous task that involves investment of substantial time, manpower, and money. Analysis of patents requires an understanding of the technology area and the product landscape, proper search methodology as well as the scalability of available tools. This is where Relecura’s Categorizer comes to your rescue. Using artificial intelligence (AI), Categorizer eliminates the manual tagging and sorting of patent data, providing you with actionable patent intelligence within minutes, which is not manually possible. It converts your training documents into AI models to better evaluate, categorize and bucket your technical data without human intervention. ‘Categorizer’ categorizes patent portfolios using a two-step process. In the first step, an AI model is created using a list of training documents. The model created in the first step is used for categorizing/bucketing patent portfolios in the second step. Multiple AI/ML/NLP algorithms available on the platform for processing technical documents are combined in this tool for providing a highly accurate categorization of patent portfolios. You can derive insights on the different sections of patent portfolios for the following actions:

• Manage patent portfolios
• Quick bucketing of thousands of documents based on training models
• Perform competitive and white-space analysis
• Map new patent portfolios to known categories
• Create training models and categorize enterprise data intelligently

How does it work?

Categorizer follows a two-step process that allows you to classify/bucket your patent data within minutes. In the first step, Model Creation, a manually-curated trained patent set is used to create an AI model that is uploaded on the platform. The model prunes the patents in the set to create a subset of the patents that are relevant to the topic. In the second step, Patent Categorization, the patents are classified via the created model using AI/ML/NLP algorithms available in the platform for processing technical documents. The accuracy is maintained by pruning out less relevant patents using the model. This provides you with categorized patent data that can be further used to gain technical insights, track technology trends and obtain market intelligence. Categorized patent data also helps you in monitoring the technology areas of your interest as well as those of your competitors. Visit Categorizer and begin making sense of your patent data. Take me to Categorizer