 |
|
|
 |
|
|
|
|
Overview
Cataphora's Technology
Much is said and written about the dramatically increasing volumes of electronic data in law firms, corporations and elsewhere. Far less attention is paid to a far more interesting and important issue that is emerging: the complexity of that data. Even as the overall size of these data collections grows (as measured by the total amount of storage space required to house them), the average size per item has been shrinking. You can find anecdotal evidence of this phenomenon by evaluating the email you receive in any one day: how much of it is two sentences in length or shorter? This is due to a combination of factors:
- The increased usage of handheld devices, such as Blackberries and Treos, which certainly discourage both the sending and receiving of longer email messages.
- The increased usage of Instant Messages (IMs) and text messaging
- The increasingly informal tone of many business conversations, as more and more people settle into a comfort zone with the use of electronic media of communication. (This is a topic that is well covered in the recent book Send)
Making things even more complicated, within the same day, any one of us might avail ourselves of any combination of email - from different accounts and devices - IMs, and text messages, not to mention the telephone. This is not because we are trying to be sneaky (at least not usually) . Rather it is because we are connected so much more of the time, and are trying to accomplish as much as possible in as little time as possible.
Such modern conveniences are all well and good until one is faced with the need to retrieve not just individual documents or records from such collections, but rather answers. For example, it is easy to perform a query with a search engine to return all documents that contain a certain word or phrase. However, if the task at hand is to determine why someone agreed to fulfill an order at a certain price that might seem unusual, the situation is considerably more difficult. This is because, unlike a tv show that starts each week by recapping the background necessary to understand the current episode, real life only rarely provides all the relevant context in a single package.
The good news is that there are likely to be many artifacts somewhere in the electronic record that collectively capture all of the person's electronically related actions for the time period in question. So there is opportunity. The bad news is that now you have to find a way to knit together the different fragments of content that will get you to an answer.
Enter the Cataphora technology, which creates the packages that are so often missing in real world communication. It alters the most fundamental assumption of existing Information Retrieval technology, which is that the individual document or record is the proper unit of analysis. Our patented discussion building engine uses the broadest possible range of evidence dimensions and sources to bind causally related items into searchable packages of data that we call discussions. It does this by creating a highly sophisticated (and usually very large and complex) model of behavior in order to help determine which items are likeliest to have a causal relationship with one another.
[Click above image to view at full size]
For example, an email might give rise to an IM response, which might motivate a phone call, which in turn might cause a transaction to occur. Even when it comes to electronic records that do not contain textual content, such as phone records, the Cataphora discussion engine is able to leverage its rich data model to tie in such items with their logical ancestors and descendants.
Discussions provide an underlying basis for all Cataphora products. In the document review platform, discussions can be used both to group and to code items together, as well as optionally for automated production. Because they are a means of imposing order and structure on unstructured and highly fragmentary data, discussions also provide an excellent mechanism for analytics that measure behavior and changes in behavior.
While discussions and related analytics are proprietary to Cataphora, we also make use of best of breed approaches throughout our products. These include:
- A highly customizable web-based user interface for both review and analysis
- A powerful and very flexible workflow system which allows customers to fully configure the rule set to be used to move documents through a traditional, automated, or mixed review.
- Three types of automated categorization approaches: ontological, supervised clustering, and unsupervised clustering.
- Choice of search engines including the powerful and very fast CQE (Cataphora Query Engine)
We also offer industry leading redaction capabilities, both for native content and images and a visualization architecture second to none for displaying analytics results.
Please Contact us for more information.
|
|
|
)
