How to Cluster to a target portfolio

How to create a clustered report of a selected organisation, and see how other portfolios read onto that

Clustering to a target portfolio

A technology cluster refers to a grouping of patent families relating to the same technical area. The platform groups patent families with similar characteristics into technology clusters, and will do so based on the patent families that have been included in your report.

In this article we discuss how you can choose to cluster one portfolio first, and see how other portfolios read onto those technology clusters e.g. if you clustered company X and are benchmarking company Y and Z: the unrelated cluster shows the number of patent families in Y and Z's portfolio that don’t cross over, or fit into, any of X’s clusters.

Lets take a look at how this works in the platform:

Benefits of clustering to a target portfolio

Clustering to a target organisation or portfolio is useful to:

• Understand which of your competitors own patents that relate to your technologies
• See what they're doing outside of these technologies too
• See where risks lie
• Perform Due Diligence - automatically compare your target patents to any organisation you select

How clustering works

Clustering is done by creating similarity matrices based on patent meta data. The clustering tool involves no human intervention or hard coded categories.

The platform uses meta-data available to create technology clusters that are as accurate as possible. Machine learning plays a role here as the algorithms identify the technology domain of the patents and give different weights to different factors (e.g. codes tend to be poor at clustering software). Meta data used to create clusters includes:

• CPC codes
• Citations (forward and backward)
• Title
• Abstract

How are cluster names determined?

Cluster names are machine generated, by reference to the title and abstract. Clusters are given a name that most closely describes all patent families in the Cluster using text summarisation, and natural language processing (NLP) techniques.

The clustering and naming algorithms are separate, ensuring that there is no possibility of a self-fulfilling prophecy in the clustering results. If the clustering were based on the occurrence of a certain phrase then it would bias that cluster towards containing only patents that used the phrase, and not other closely related technologies irrespective of words, skewing the clustering results.

How do you determine a cut off for the number of clusters in report as this could be endless?

The platform will present a maximum of sixteen technology clusters. If a group of patent families creates more than sixteen clusters, the clustering tool will group the remaining families into ‘miscellaneous’.

If relevant to your reporting, this miscellaneous cluster can then be broken down into further clusters for your analysis.

Clustering to a target portfolio - Step by Step image walk through

1. Click 'Start' on the new report tile:

2. Search for the target portfolio and additional portfolios you want to add to your report. Click to add to your report basket:

3. Once you've added all required portfolios to the report basket, select 'Next' without applying any classifiers.

4. On the next screen clustering is selected for all portfolios by default. Ensure that the only tick that remains in the clustering options is for the target organisation that you want to cluster (group) first. In this instance I've chosen Ford:

5. Select 'Run' to start your report. The Ford portfolio will be clustered first, and then we'll be able to see which assets for GM, Mercedes and Volkswagen fall into these groups. There will also be an 'unrelated' cluster for these three orgs that don't fall into the Ford cluster categories.