Categorizing patent data for finding gaps and opportunities

Categorizing patent data
for finding gaps and
opportunities

Finding gaps in the technology space (also known as white space analysis) provides your legal and technology teams with inputs on where there is room for technology development and protection and thereby tap a technology area ahead of the competition. The two common reasons to carry out a white space analysis include:

Expansion into new technology areas or enhancing the existing technology and product portfolio. The process can also yield adjacent opportunities that complement your existing portfolio.
Filling up gaps in your existing technology-patent mapping by plotting your existing IP against a Problem-vs-Solutions Matrix and finding out areas that you haven’t fenced yet.

Why do you need to categorize your patents?

Patents disclose information in complex language and in a complex manner. Extracting the right industry-specific concepts and information from patents is crucial for most analyses. This is where categorization of patents comes into action. Categorization is necessary to bring real-world business context to various parameters involved in the patent document. It is done to align patent datasets to the way your company understands the business. You can think of this as being able to see the patents with your own lens.

How to decide what categories to choose for your patents?

The same patent data can be categorized in many ways. The choice of categories makes a real impact on your analysis and so this “thinking” step is critical. Some of the common styles of categories are Problems, Solutions to the problems, Product parts, processes used, compositions, ingredients, or materials. Platforms such as PatSeer can also extract technology terms from the patent data which could be a raw set of technology terms (keywords) or more refined topics. So, you can also map the Problems with the auto-generated keywords/Topics to see if alternative technologies can be mapped to the same problem.

It’s important to look into your current business and market data to identify the problems that exist and the known set of solutions or methods or compositions against those problems. Proper planning before doing categorization will help you go in the right direction.

Bucketing the patents into the categories

The methods to categorize fall into three basic types namely:

Manual categorization – Manually going through each record and placing it into a category gives you the highest precision but at the cost of time and effort.
Semi-automated categorization – Use specialized search queries to filter your patents and push them into a category. This process has a fairly high precision if the queries are made by search experts.
Fully automated categorization – Use of tools to autofill and automatically add patents to your categories. By leveraging a combination of AI and NLP this method has been seeing acceptable levels of precision in recent times.

PatSeer’s autofill categories option automatically adds patents to categories based on keywords, classification codes and concepts in the main body of the patent. This feature helps you save valuable time and bucketize large patent datasets quickly.

Sample patent categorization use case

Let us take an example to study the categorization and mapping of patents for better understanding.

Here we have taken Hand Gesture Recognition Systems as the example technology space. We have categorized the dataset against technologies/sensors used, by applications and by input device types. (See below)

Once the patents were bucketized into these categories, various matrix analyses were performed to identify possible gaps.

As seen in the Technology vs Application matrix above, various types of sensors have been mapped to applications and we can easily see areas where certain sensors have not been utilized. These can be explored further.

Similarly, a heatmap of Sensors vs Input device types shows a gap in use of proximity sensors, capacitive sensors in virtual reality gloves and use of photoelectric sensors and RGB-D sensors in cyber glove and virtual reality glove. Such gaps can act as potential white spaces and should be analyzed further.

Categorization of patents is the foundation for doing gap analysis. Identified gaps should be discussed with technical experts to evaluate the feasibility of converting them into opportunities.

Read more about white space analysis– ‘Taming Patent White-space analysis – a guide for IP owners’

PatSeer contains a complete set of pre-processing, analysis, and mapping tools to aid in each step of finding gaps. Create clusters/topic maps, IPC/CPC based maps, create custom taxonomy and mapping patents to create categories. Use matrix-based analysis to analyze category buckets and find white spaces, if any. Multi-generation forward citation analysis feature will help you track the use of technology in different application areas and to track competitor’s patent filing trends.