|
The Deep Query Manager™ is an extremely capable and functional document
harvesting, tracking, management, analysis, reporting, collaboration,
results and publication platform. This page, and its two accompanying
siblings, presents a few of the DQM's screens among the multitude
available in the product. This page covers Projects and Reports. Across all three related pages the sections include:
Projects
Projects are the central organizing construct for documents within the
Deep Query Manager™. Multiple harvests from multiple source groups can
be combined into the logical work unit of the project. For example,
each of your major competitors could be assigned to their own named
project, with various harvests or probes contributing, say, financial
documents, technical documents, sales documents or press commentary as
one single analytic unit. Projects can be split, combined and data
mined. Projects are often used as stages within analytic pipelines
where the vetted results from one step in the knowledge management
process become input to the next stage.
The next screen shows how multiple individual harvests can
contribute to a single project. It also shows that the harvests
themselves can be combined or manipulated in various ways. As is
typical with most DQM results views, a variety of sorting,
presentation, slicing-and-dicing and export or sharing options are
constantly available.
Any subset of results can itself also be locally searched, exported
or used to create a new project or results set. Within any given
document subset, all documents are scored on an equivalent basis with
all duplicates removed. Dynamic summaries are created based on the
current local query (or in its absence, the original retrieval query).
Groups and range select options are available, as well as viewing the
document in its original downloaded form (as would be displayed on the
Internet) or in its dehydrated text form as contained within the index.
Any and all of the Deep Query Manager™ features can be applied to up to 140 different foreign languages. A similar results display is shown below for Arabic:
One or more document results, up to and including whatever results
set is currently active, can have a terms list generated. This feature
is useful for refining queries or as an aid in creating new dictionary
filters. (For example, false positive results can be selectively
analyzed to create an exclusion filter.) As with all other objects in
the DQM system, new filter groups can be named and combined with other
groups.
Back to top
Reports
Of course, the generally desired outcome from this purposeful
harvesting, organization and refining of results is to create reports
to colleagues or management, or for submission to other analytics. The
Deep Query Manager™ combines the concepts of a template or view (a display format) with qualification filters and restrictions (such as queries or metadata results), applied to a vetted set of results
(project or other subset) to create a report for presentation or
export. Some of these templates are for presentation or visualization
purposes; others are for export and automatic data processing purposes
(such as XML). Furthermore, once these various report configurations
have been developed, they can be named and reused with still other
results sets or projects, or exported to colleagues for application to
their own data.
Metadata processing, indexing and support is still another strength
of the Deep Query Manager™. The DQM can recognize and process
literally millions of different metadata tags, with system
administrators able to set weighting factors for tag presences or
absences. Local searches and data mining can combine or not any
metadata with full-body text.
The Comparison Report is especially geared to monitoring and tracking, though any two results sets can be compared:
In comparison mode, the documents from both candidate sets are placed into one of four bins: new, modified, absent or unchanged. The new
view is helpful when tracking or monitoring sites. In practice, DQM
users gain a one to two order magnitude productivity improvement in
site monitoring over any other technique. The absent
(deleted) view can be helpful to point out embarassing content removed
from a site (the DQM is often used for political candidate tracking in
this view mode). In the case below, we are viewing the documents that
have been modified in the time interval between the baseline and comparison results sets:
Modified documents have a special report wherein the actual text
changes in the before and after cases can be shown, with the
differences highlighted by color. The example below is for Arabic,
again showing that all Deep Query Manager™ functionality can be applied
to any supported language:
Among the variety of still other specialty reports is the example
below showing retrieval acceptance success for documents downloaded
from multiple searchable databases, retrieved simultaneously for a
given harvest. This report is helpful to determine great-producing
source groups for repeated harvesting purposes, among other uses.
Back to top
|
