The technology cluster to which the patent family belongs. The technology clusters are technology fields created and updated by PatentSight and have four hierarchical levels (e.g., "Consumer > Sports > Ball game > Ball feeding"). In contrast to IPC or CPC, each patent family can only belong to one technology cluster per level. The clustering process is performed weekly with each database update, thereby incorporating new publications into the technology clusters. All patents, patent applications, and other IP rights in the PatentSight database are assigned to one cluster on each level.
PatentSight’s technology clusters are sets of similar patent families created based on full-text patent data under consideration of IPC classifications by an automated machine learning technique. The technology clusters are built based on all documents from all patent families and other IP rights in the PatentSight database irrespective of their legal status and have four levels of hierarchy.
- First, PatentSight applies a self-developed IPC mapping process to allocate each patent document to one or more of 135 level 2 clusters. At this stage of the clustering process, the patent documents can be mapped to multiple level 2 clusters.
- The level 2 clusters serve as a basis for creating the level 4 clusters, which are created based on full-text data. The documents in each level 2 cluster are processed into word stems, including titles, abstracts, claims, and descriptions. The word similarity is measured in the semantic context of each level 2 cluster. Taking into consideration the frequency of these word stems, the stems are mapped to vectors and clustered, thereby resulting in the level 4 clusters.
- Then, the level 4 clusters are automatically aggregated into level 3 clusters. This process is also based on full-text data.
- Next, PatentSight manually aggregates the level 2 clusters into a total of 13 level 1 clusters.
- In the following step, the patent documents are assigned to the most fitting level 4 cluster. Based on its level 4 cluster, each patent document is also assigned to the overlying level 3, 2, and 1 cluster. While in the first stage of the clustering process patent documents can be mapped to multiple of the 135 technologies, this is no longer the case after the assignment of the patent documents to the clusters (i.e., there is no co-classification).
- Finally, the clusters are automatically labelled based on the titles and abstracts of the patent documents.
Freedom of Overlaps
Since each patent family in the PatentSight database belongs to exactly one technology cluster on each level, portfolios can be analyzed without having to pay attention to technology overlaps (e.g., as in IPC classifications).
The weekly inflow of new documents into the PatentSight database results in continuous updates of the technology clusters. These weekly updates may come with changes to the overall number of level 3 and 4 clusters (with an increasing tendency), the reallocation of patent families to different clusters, and changes to the cluster labels.
Based on iterative training, PatentSight’s clustering approach allows for easy incorporation of new data. The automated machine learning continuously improves the clustering process.
How to Use Technology Clusters
You can use the search field "Technology Cluster" in the search panel to narrow down your search results to patent families that belong to a technology field of interest.
Searching by Technology clusters can be helpful in prior art searches, FTO searches, benchmarks, and many more use cases!
Click on the "More" icon to access all available search fields.
|Access the section "Technology". Alternatively, enter "Technology" into the search box of the "Add Search Field" dialog.
|Select the search field "Technology Cluster."
Click on the drop-down icon displayed next to a cluster of interest to access the lower levels.
Tick the checkboxes of the relevant technology clusters you want to add to the search panel.
After narrowing down your search results to patent families that belong to a technology cluster of interest, you can for example find out which owner owns portfolios in this technology field and how these portfolios have developed over time.
This example line chart shows that, in the field of "Circuit boards" (as defined in the search query), Company B owned the strongest portfolio as of Patent Asset Index until 2021 and was then overtaken by Company A.
Alternatively, you can use the sunburst chart to obtain an overview of an owner’s portfolio grouped by technology clusters.
Click "Insert" in the main tool bar and select "Chart." In the chart settings (accessible via the gear wheel icon), select the chart type "Sunburst."
|Double - Click on a segment displayed in the sunburst chart to expand the segment. Double - click on an expanded segment to collapse it again.
Click on the pink arrow to display hidden segment.
Under "Segments" you can slice the sunburst chart by the one of the attributes "Technology Clusters" or "IPC (Smart)—Short Title."
|Under "Segment Size", you can select the measure (here: “Portfolio Size”) to be visualized by the size of each segment.
|Under "Technology Cluster Levels Shown", you can move the slider to adjust the number of technology cluster levels displayed in the sunburst chart.
|Use the radio buttons under "Level of Details" to adjust the level of details shown in the sunburst chart. The setting "Low" will only show the main technology clusters, whereas the setting "High" will show a higher number of sub-levels.
You can use the attribute "Technology Clusters" or "Technology Cluster (Owner and Title)" and the measure "Share of Parameter Selected in Technology Cluster" in all charts and tables available in PatentSight.
In this example, the table is grouped by the attribute "Technology Cluster—Level 3 (Path and Title)."
|The table shows the measure "Share in Total Number of Active Families in Technology Clusters (Level 3)" to find out in which technology an owner occupies a strong position.
|The owner Company A owns 35.7% of all active patents belonging to the level 3 technology cluster "Surface acoustic wave."
Finally, you can also use the technology clusters to quickly find out in which technology areas competitors are active and own strong patent portfolios. This example matrix chart shows that Company C has the strongest portfolio in the field of "Automatic Transmission."