Fortune Interactive Corporate Blog
01/26/06: The SEMasphere™ - a mini white paper
So what's this SEMasphere™ all about and why is it important?
The SEMasphere™ Interface is a new module in SEMLogic™ that is a virtual reality (VR) environment. This VR module enables our technicians to see important relationships within data and intelligence in ways not possible before – making attainable the merging of human judgment, intuition, and expertise with the artificial intelligence already a central component of the SEMLogic™ system. It builds on theory from information design, computer graphics, human-computer interaction and cognitive science.
As the competitive intelligence information continues to increase from our research, it is important that the correct information is being seen by the right people in an understandable format. The ability to view graphic representations of competitive intelligence data empowers us to make better marketing decisions quickly.
The SEMasphere™ is essentially visual analytics, the formation of abstract visual metaphors in combination with a human information discourse (interaction) that enables detection of the expected and discovery of the unexpected within massive, dynamically changing information spaces. Beyond the mere transfer of facts, the SEMasphere™ aims to further transfer insights, experiences, expectations, perspectives, and predictions by merging various complementary visualizations.
A SEMasphere™ is created for a particular keyphrase we research. Each SEMasphere™ is based on an analysis of over 100,000 data points from off and on-page factors of various web pages. This underlying analysis also incorporates multiple levels of artificial intelligence and quantum programming.
Based on multiple vector representations (similar to what is depicted in Figure 1.) in n-dimensional spaces of the data collected, we are able to apply several dimensionality reduction techniques (such as those depicted in Figures 2 and 3.) to reduce the space to three dimensions so that it can be viewed. (And just in case you were wondering, I'm not just talking about term vector databases - this goes far beyond that.) The view seen in the SEMasphere™ is actually a blend of multiple disparate n-dimensional spaces.
Figure 1. Vector Representation
Figure 2. Principal Component Analysis - one standard technique for dimensionality reduction
Figure 3. Singular Value Decomposition - another standard technique for dimensionality reduction
Bottom line: This allows us to take enormous amounts of important data, shape that into relevant information, and convert that to actionable intelligence which can be seen visually and manipulated dynamically.
: : Demonstration
Figure 4. Portion of a SEMasphere™
Figure 4 is a small portion of a SEMasphere™ which shows prominently three spheres (2 red, 1 blue). These spheres represent three different web pages. Their close proximity to each other in the VR environment indicates that these pages are similar to one another with respect to many of the features being analyzed. However, the difference in their appearance also indicates some dissimilarities. The question is whether these are differences that make a difference.
In every SEMasphere™ there is a zone which demarcates the spheres that are revealed statistically to be very important. This zone is a high dimensional version of a cube, a hypercube. We can determine the strengths and weaknesses of any web page by the position of its sphere with respect to the hypercube: whether its inside or outside the hypercube and how far it is (in n-dimensional space) from the hypercube. Recall that this is determined by analyzing over 100,000 data points.
Everything outside the hypercube is segmented into three zones, concentric regions with various distances from the hypercube. The three zones are indicated in the SEMasphere™ by stripes on a sphere. The more stripes (from 1 to 3), the further away it is from the hypercube, the poorer the quality of the page represented.
Figure 4 shows one red sphere with 3 stripes, one red sphere with 1 stripe, and one blue sphere with no stripes. We can determine the relative quality of the pages with regard to a multitude of on and off-page factors this way. Although their relatively close proximity to one another indicates important similarities, we know that their dissimilarities are differences that make a difference. Clicking on a sphere would actually provide a host of detailed data for that page from which we can determine what it is that made the difference.
We know how high or how poor the quality is of any page analyzed, its strengths and weaknesses, and what needs to be done to improve that page based on over 100,000 data points.
"The ability to learn faster than your competitors may be the only sustainable competitive advantage."
- Arie P. De Geus, former coordinator, group planning, Royal Dutch/Shell quoted by Peter M. Senge (source: "The Fifth Discipline" )
[Mail to a friend]
: : So much data; so little time.
The SEMasphere™ Interface is a new module in SEMLogic™ that is a virtual reality (VR) environment. This VR module enables our technicians to see important relationships within data and intelligence in ways not possible before – making attainable the merging of human judgment, intuition, and expertise with the artificial intelligence already a central component of the SEMLogic™ system. It builds on theory from information design, computer graphics, human-computer interaction and cognitive science.
As the competitive intelligence information continues to increase from our research, it is important that the correct information is being seen by the right people in an understandable format. The ability to view graphic representations of competitive intelligence data empowers us to make better marketing decisions quickly.
The SEMasphere™ is essentially visual analytics, the formation of abstract visual metaphors in combination with a human information discourse (interaction) that enables detection of the expected and discovery of the unexpected within massive, dynamically changing information spaces. Beyond the mere transfer of facts, the SEMasphere™ aims to further transfer insights, experiences, expectations, perspectives, and predictions by merging various complementary visualizations.
: : A visualization is worth a 100,000 data points.
A SEMasphere™ is created for a particular keyphrase we research. Each SEMasphere™ is based on an analysis of over 100,000 data points from off and on-page factors of various web pages. This underlying analysis also incorporates multiple levels of artificial intelligence and quantum programming.
Based on multiple vector representations (similar to what is depicted in Figure 1.) in n-dimensional spaces of the data collected, we are able to apply several dimensionality reduction techniques (such as those depicted in Figures 2 and 3.) to reduce the space to three dimensions so that it can be viewed. (And just in case you were wondering, I'm not just talking about term vector databases - this goes far beyond that.) The view seen in the SEMasphere™ is actually a blend of multiple disparate n-dimensional spaces.
Figure 1. Vector Representation
Figure 2. Principal Component Analysis - one standard technique for dimensionality reduction
Figure 3. Singular Value Decomposition - another standard technique for dimensionality reduction
Bottom line: This allows us to take enormous amounts of important data, shape that into relevant information, and convert that to actionable intelligence which can be seen visually and manipulated dynamically.
: : DemonstrationWitness the instructive power
of a
fully functional SEMasphere™.
of a
fully functional SEMasphere™.
Figure 4. Portion of a SEMasphere™
Figure 4 is a small portion of a SEMasphere™ which shows prominently three spheres (2 red, 1 blue). These spheres represent three different web pages. Their close proximity to each other in the VR environment indicates that these pages are similar to one another with respect to many of the features being analyzed. However, the difference in their appearance also indicates some dissimilarities. The question is whether these are differences that make a difference.
In every SEMasphere™ there is a zone which demarcates the spheres that are revealed statistically to be very important. This zone is a high dimensional version of a cube, a hypercube. We can determine the strengths and weaknesses of any web page by the position of its sphere with respect to the hypercube: whether its inside or outside the hypercube and how far it is (in n-dimensional space) from the hypercube. Recall that this is determined by analyzing over 100,000 data points.
Everything outside the hypercube is segmented into three zones, concentric regions with various distances from the hypercube. The three zones are indicated in the SEMasphere™ by stripes on a sphere. The more stripes (from 1 to 3), the further away it is from the hypercube, the poorer the quality of the page represented.
Figure 4 shows one red sphere with 3 stripes, one red sphere with 1 stripe, and one blue sphere with no stripes. We can determine the relative quality of the pages with regard to a multitude of on and off-page factors this way. Although their relatively close proximity to one another indicates important similarities, we know that their dissimilarities are differences that make a difference. Clicking on a sphere would actually provide a host of detailed data for that page from which we can determine what it is that made the difference.
We know how high or how poor the quality is of any page analyzed, its strengths and weaknesses, and what needs to be done to improve that page based on over 100,000 data points.
- What would normally take days of analyzing dozens of pages of detail data, now takes only moments!
- What was once only discernable by a best guess (who can ponder over 100,000 interconnected data points and their permutations in their head at one time?), now is known with great precision - no guess work!
- We know how well or poorly optimized a page is for on and off-page factors and why.
- We enable our clients to learn more about their competition, more accurately, and empower them to take action with confidence.
"The ability to learn faster than your competitors may be the only sustainable competitive advantage."
- Arie P. De Geus, former coordinator, group planning, Royal Dutch/Shell quoted by Peter M. Senge (source: "The Fifth Discipline" )
[Mail to a friend]
Category: SEMLogic - [Printer friendly version]
|
Posted by: Mike
|
Want a blog like this? Check out our corporate blogging services.
Want a blog like this? Check out our corporate blogging services.
Comments
No comments yet