Halitus Patent Cluster Analysis

Goal: Getting a fast and precise technology impact evaluation of Halitus` inventions.


  1. We received three Halitus invention concepts from which we used 3 problem-descriptive paragraphs describing what shall be solved and in which market segment it would be needed, as well as three short descriptions of  solutions for the problem.
  2. We have built a full technology field bottom-up based on the six concepts, containing a large technological space with 10.000 patent families, wherein we will be able to locate a huge majority of case relevant  patents, players and concepts.
  3. We have built digital twins of the six concepts and projected these into the field, as well.
  4. We divided the field into 50 clusters and identified relevant clusters by visualizing appearance of the digital twins. The problem-related concepts were found in one breath analysis cluster (green), while the solution concepts were found in gas analysis clusters farther away (orange).
  5. The breath analysis cluster is separated from the field and includes a very precise collection of patents by competitors, potential partners, similar or different technical solutions as well as other but strong related business alternatives (to be analysed in detail together with customer).

Result: Halitus’ inventions target the field of breath analysis. They identified a technical problem to be solved, which seems to be rather unique compared to the general idea of the field (not found in any cluster, which represent >85% of the technical field). Halitus retrieved inventive solutions to the problem from other gas-analysis related areas and have bridged the problem space with a new solution space (optical sensors, cartridge mode for variability). From a first analysis we can see at least three major players in the final market (Philips, Bosch, Resmed, maybe also Roche) as well as several smaller players with somewhat different technical solutions and business fields. The target market field shows a certain dynamic of activity, but is not exploding yet. Patents of high quality are mixed with lower scores, which indicates a not yet fully explored field.  A closer analysis of the funding situation of these smaller players might indicate money flow as well as expectations. The retrieved patents should be well suited to produce a fast evaluation of the competitive environment and FTO-situation for the main problem target. Several other business areas, such as automotive, fitness, wellness, etc. are located which can be explored quickly from the result list.

Cluster Distance Analysis

In contrast to older 2D representations of content-distance, we define clusters of conceptual closeness different: In our multi-label approach we can measure how many patents are not only in one but in several clusters, because their texts contain more than one concept. Typically, problems and solutions are not exactly in the same concept space, but have a relevant closeness to concepts in other clusters. This can be expressed by a closeness factor. From the perspective of Halitus’ target market cluster breath analysis (green) we can identify other spectral gas analysis technologies, such as biogas- and chemical reaction gases or industrial solutions, home- and pollution gas monitoring, transport security (for insurances or worker safety) or semiconductor production, all within a certain distance. Clusters far away from the target cluster do not have any patent in common. The same approach as demonstrated here for the green target cluster can be applied to all other clusters, to visualize conceptual closeness to any target.