Main    Products    Solutions    Technology    Pricing    Partners    Support    Corporate Info    Feedback    Contact Us
Products :

Core Systems

Hosted Online Solutions

PatternScape:


Pattern profiling search, cluster views and visualization
PS is a pattern profiling search technology in which information landscapes are mapped and clustered according to patterns, traits, markers and features. Search results are formed into clusters that are easy to navigate, and comprehend, regardless of the complexity of the information being mapped. The system employs a high speed, inverted index, search framework. This framework renders high-speed alphanumeric results to the PS clustering engine, which determines visualization and inter-relating results.

PatternScape is useful for search and analysis of inter-related information. It employs a mathematical model the result of which is a visually comprehensive presentation of clustered search results. This intrinsic capability to cluster information is what makes possible the notion of distinction and user-determined-choice during search. Complicated though it might be under the hood, PS is a simple to use system for information visualization, occurrence, frequency and co-existence clustering.

The system can be used in a variety of ways, a few examples are:
  • General documentation search (information repository)
  • Proprietary information search (knowledge repository)
  • Specific analytic search (intelligence repository)
  • ATM/ABM debit card fraud tracking by mapping card use with co-existing previous event analysis. see brochures.
  • Insurance underwriting mapping and analysis for determining insurability based on historical records search and analysis.
  • Mortgage application mapping and analysis with previous event to uncover fraud or viability of application
  • Mapping information landscape
  • Market opportunity analysis and macro-micro assessment
  • Investment portfolio and corporate business development analysis
  • Criminal profiling, crime scene evidence co-relation, and trait mapping
  • Prospecting opportunity analysis, assay evaluations, and cross-comparisons
  • Medical symptom clustering and disease analysis
  • DNA base pair sequence, co-existence, and clustering analysis
  • Marketing analysis for product differentiation and comparison to competitor voids
  • Sale prospect analysis, competition analysis, as well as client needs profiling
  • Engineering design analysis and project information archiving, and analysis
  • Resume and job type mapping
Though the system can be used as a search and find tool for information, documents, or data research, its ultimate performance can be found where variants, information distinction and specific choice is needed by the professional user.

Most importantly, the system is a simple to use visual tool even though based on complex hidden processes. The table shown describes configurations and options of each main feature in the framework such as DocMap, Co-x, 2x2 search pad, weights and synonym variants. Screenshots are provided for review.

PatternScape 2x2 3x3 4x4 5x5 Cluster
View		 Link
View		 Vectors
Weights		 Synonyms User
Mgmt		 Book-
marks		 Doc-
Map		 Co-X Multi-
Target
Index		 Multi-
Source
Index
PS-1 x       x x              
PS-2 x       x x   x x        
PS-3 x       x x x x x     x
PS-4 x     x x x x x x   x
PS-Enterprise x x x x x x x x x x   x
PS-Extreme x x x x x x x x x x x x x x

Features:

2x2 matrix search pad
The search pad is used to enter different words and phrases from which a ClusterView tree appears. This tree and branch system shows variants for all four terms (phrase) entered into the entry pad. Technically speaking this is a four-root variant map from which the user may navigate to results (LinkView). LinkView contains the documents that are comprised of the 2x2 terms-all this in a cluster distribution for pertinent results analysis. The system provides result variants based on the users search needs. These variants render observations of the information landscape contained in vast quantities of documents and data. These vast quantities of search results are reduced to observable clusters with analytic distinction.

      

3x3, 4x4, 5x5 matrix enhancements
Enhanced search-pads provide significant increases in clustering granularity during search. Document Map and Co-existence fidelity (DocMap and Co-x) impacts search with this increased cluster granularity. Cluster granularity refines variants, during search, which leads to user controlled information mining and distinction.

Weights within 2x2
The weights option is a facility for users to ascribe different impact values to each entry in the search pad. For example, the researcher might over-weigh "gold inclusion" as 100%. This will ensure that results of "gold inclusion" will be biased in finding results. You can also imagine how weights can help an officer render order-importance in criminal records. PS puts the order of importance of results in the users hands with features like weights and ClusterView. Vectors and Matrices further intensify this notion, as discussed later in the page.

Example of weights as an important distinguishing tool
Weights are used for applying importance or power values to the individual elements of the search pad matrix. For example, suppose you where analyzing eight and twelve-cylinder engines where torque was more important than power in your search. The user may increase the weight of the torque element-the effect of which is cluster distribution vectors that place torque-containing documents with a distinguishing position in the results. Boost is applied to vector distribution across the entire space. DNA cluster analysis might well be another example of its use for preponderance of pairs, markers of gene replications or mutations, for instance.

Co-existence mapping (Co-x)
Co-x usage determines the co-existing inter-relationship of alphanumeric terms and phrases, in cluster. 

Example:  “sore throat” will render a ClusterView of co-existing terms with all known combinations. The result is a complex inter-relating cluster map, which allows the user to select the ideal result that can be submitted to the 2x2 - 5x5 engine for document and data retrieval.

Co-x is ideal for helping the user ascertain co-existence and inter-relationship. It is curious to find that "sore throat" and "acute pharyngitis" co-exist in a highly predominant way throughout an index while "Vicodin Lortab" and "pain relief" also inter-relate, for example. Co-x operates across the entire target index space regardless of the complexity, or simplicity, of inter-relationships.

      

Document mapping (DocMap)
Contrasted with co-x, DocMap is similar operates across a single document only (or URL). For example, it can be used to map ones resume, submit resulting skills-clusters to the employment index to find combinations of matching job assignments. Matching results can also apply to a medical assessment, a job specification, or a criminal record.

The 2x2 search pad is auto-populated by the DocMap result. Click on resume cluster branches, for example, and the 2x2 pad will auto-populate for pattern search against the entire job index, thereby finding matching job assignments (similar for mapping of diseases, criminals, investments, or DNA pairs). Regardless of the vastness of the document repository, DocMap finds inter-related results with the input document or URL.

Cluster views (ClusterView)
Cluster views, the variant maps created from the matrix search pad, contain valuable information about frequency, structure and occurrence in the data set and their vector inter-relationships. Vectors map complex relationships of terms, mathematically, with one another throughout the document space. The resulting tree structure of relationships helps users visualize and navigate to relevant information. ClusterViews are at the core of PS intensified search.

Link views (LinkView)
Link views are the resulting documents relating to a ClusterView branch (they are presented as classical web site search engine result pages). As a distinction from classical results, however, LinkView presents vector distribution and power values for document-to-document comparison. The predominance of one term over another is illustrated with vector distributions. These distribution vectors exhibit predominance of traits, evidence, symptoms and markers across the resulting group of documents. Thus PS empowers the user to evaluate search results with visualization clusters and not  endless result lists.

    

Distribution Vectors (eVectors)
A distribution vector is used to evaluate the distribution of frequencies-of-occurrence of terms in cluster. The vector is an extremely important metric, an example of which is in mining:  inter-relating  distribution of compounds in an ore sample, for example.

Distribution vectors "appear" useful for only the most deliberate and professional user but user friendliness provides insight for simple search matters such as engine power and torque distribution, for example. Another simple example, yet important to the traveler:  seaside views versus golf courses. Whatever your use of PS, distribution vectors enhance information distinctiveness and visualization.

eValues
An eValue is a power value ascribed to the distribution vector and it illustrates cluster size or simply stated the importance of a search query compared with its co-relating neighbors. The cluster size of one over another will often imply power of one set of findings over another. The eValue and distribution vector share a courtship of meaning to the individual needs of the user. Together with weights they form a method for the user to control analysis and visualize distinctiveness.

Document and hit frequencies
Hits are frequency-of-occurrence search terms across a cluster, an index (space). Hits are normally shown in search engines as the number of documents, which contain the search term. In PatternScape, hits and document frequency are similar with the following distinction: cluster phrase occurrence impact, power and elemental distribution.

A-Matrix for advanced data set analysis
The PatternScape A-matrix follows subspace math principles for finding a spatial distribution (characteristic) matrix:

Synonyms
PS handles synonyms in a variety of unusual ways: In SoftPaperScape and SpeakScape, synonyms are error detection and repair tools. In 2x2 matrix search synonyms are employed by the user for correct phrase and term choice. The system presents synonyms for the terms entered in the search pad, which can be used to refine search with synonym replacement. Synonyms can also be used for alphanumeric purposes like financial, engineering and microbiology, for instance. Thus user defined synonym lists are possible.

Multiple target indexes and x-Search sites
PS allows for selecting one of many different indexes in which to apply search. The user might well have a crime index and a forensic science index configured in their system. Rather than apply search to both indexes the user may choose to examine results only found in the crime index. It is easy to imagine a clinical scientist having many indexes on their system for various application and analysis (multi-Genome). Comparatively, a journalist might have two different information resources to apply search. It is a powerful tool for the specialist to be able to flip flop between information assets (indexes). And CrawlScape provides the automated framework for quickly creating and managing these assets.

Multiple target indexes and xSearch sites
You might use your own index to find clusters but wish these be produced across a different target database for actual document results. This is how one would use multiple targets: For example an FBI agent searches the criminal database and then applies resulting clusters to the Interpol international index for equivalent comparison.

Bookmarks and Member Management
Bookmarks and member management allow for users to identify themselves and save bookmarks of their various Search, DocMap and Co-x pads. The option makes sure serious analytic search projects are saved and retrievable.

Exporting LinkView Results, ClusterView Maps, and Bookmarks
PS allows for exporting valuable metrics (vectors, links, clusters, (f), and power values) so users can evaluate in external information manipulation tools.

Copyright © 2006 KnowledgeShape. All rights reserved.