KB484307: Pattern Matching vs. Keyword Matching

Product Manager, Senior • Strategy

HyperIntelligence allows you to add various, or alternate, forms of the keyword attribute for matching. In this article, we clarify your questions regarding Pattern Matching and how it is different from the standard Keyword Matching. HyperWeb surfaces cards on a web browser based on a keyword that matches the underlying criteria set by the card author. These criteria could be either a simple text-based keyword or a pattern. So, let’s look at which approach best suits your business needs and how the end-user experience looks like!

What is Pattern Matching?

As part of the Strategy 2020 Update 2 release, you can now match keywords based on a character pattern. This enables analysts to accommodate use cases with larger datasets, or card sets, in both HyperIntelligence for Web and HyperIntelligence for Office. With this new technique called Pattern Matching, you can avoid downloading (or caching) the keywords in the browser storage for better and faster performance.

If your dataset or card set consists of objects that include a pattern and you have identified that object to be your primary keyword for matching, then you should consider enabling pattern matching for that card set.
Here are some examples of real-world data that have strong patterns:

SBN Code (Publishing, Media)
Example: 978-3-16-148410-0

SKU (Retail, Sales)
Example: AE-SK-JN-34-DT

VIN (Transportation, Logistics)
Example: 1C3BF66P0GX570598

Policy numbers (Insurance, Regulations, Government)
Example: 99999999-1

Bank account numbers (Finance, Banking)
Example: 01234567890123

What is Keyword Matching?

To use HyperIntelligence, the card author chooses a keyword (or an object from the dataset) and its alternate forms to enable matching on the client. These keywords may or may not have a pattern. In most cases, the keywords are product names, employee names, customer names, and so on. The card author doesn't have to take any extra step in enabling keyword matching since it's set by default and uses the selection under alternate keyword matching section.

Here are some examples of data that do not have strong patterns:

Employee Name
Example: John Smith or Smith, John

Customer Name
Example: Spencer and Kutch Inc.

In summary, we recommend you use pattern matching if you have a large dataset (greater than 500K keywords), and the object you want to match on consists of a pattern.
The table below summarizes the differences between the two types of matching:

Pattern Matching	Keyword Matching
Supports datasets with more than 500,000 unique keywords.	Supports datasets up to 500,000 unique keywords.
HyperWeb and HyperIntelligence for Office looks for the unique character pattern in the keywords and surfaces cards based on that criteria.	HyperWeb and HyperIntelligence for Office looks for an exact match of the keyword that is stored in the dataset.
Card keywords are not downloaded (or cached) in the browser storage.	Card keywords are downloaded (or cached) in the browser storage.
You will not be able to use the search functionality to explore pattern enabled card sets	You will be able to use the search functionality to explore non-pattern enabled card sets.

To learn more on how to enable pattern matching, view our product documentation.

Comment

0 comments

Details

Knowledge Article

Published:

June 9, 2020

Last Updated:

June 24, 2020

What is Pattern Matching?

SBN Code (Publishing, Media)
Example: 978-3-16-148410-0

SKU (Retail, Sales)
Example: AE-SK-JN-34-DT

VIN (Transportation, Logistics)
Example: 1C3BF66P0GX570598

Policy numbers (Insurance, Regulations, Government)
Example: 99999999-1

Bank account numbers (Finance, Banking)
Example: 01234567890123

What is Keyword Matching?

Employee Name
Example: John Smith or Smith, John

Customer Name
Example: Spencer and Kutch Inc.

Pattern Matching	Keyword Matching
Supports datasets with more than 500,000 unique keywords.	Supports datasets up to 500,000 unique keywords.
HyperWeb and HyperIntelligence for Office looks for the unique character pattern in the keywords and surfaces cards based on that criteria.	HyperWeb and HyperIntelligence for Office looks for an exact match of the keyword that is stored in the dataset.
Card keywords are not downloaded (or cached) in the browser storage.	Card keywords are downloaded (or cached) in the browser storage.
You will not be able to use the search functionality to explore pattern enabled card sets	You will be able to use the search functionality to explore non-pattern enabled card sets.

To learn more on how to enable pattern matching, view our product documentation.