The statistics so often are wrong or, at least, are wrong their users.
Are wrong in applying the tools of statistical science putting together
fragments of informational objects or situations one from each other,
completely different.
First they fragment, then mix and finally aggregate.
Finally they get a conclusion.
So the researchers on political trends break up the views of
respondents, mix the individual responses, aggregate, and finally claim
truth that are attributable only to the respondents that they have
created in virtual way, non-existing in reality and certainly not
attributable to individuals or homogeneous groups of respondents.
Similarly, the Business Intelligence makes available the tools of data
analysis; it can cut the data and then reassembling it into
multidimensional structures in which the peculiarities of information
starting positions have been destroyed.
So the Business Intelligence brings together companies from different
sectors, revenues, organic markets and payment habits, changing from
time to time variables crossing data.
Which subject (or situations) may be applied then the decisions are
taken, having destroyed the wealth of information overall subject (or
situations) starts from?
For example, if you had an archive of animals including mammals, in
which there were also men and primates, as a result I could get that
mammals have on average about three feet.
Where can I find a mammal that has an average of three feet?
To make real statistics it's necessary to be maintained intact as much as
possible the information assets of the initial data of the subject or
the situation under review.
The techniques derived from Neural Networks use an approach to data
analysis entirely respectful of the information assets of the initial data.
In fact do not require the user to define variables to cross, thereby
preventing him from making strange crosses.
Only require you to enter the maximum number of groups that the
algorithm will create.
Do not destroy the wealth of information of the initial data, but work
always subject's data (or situation) in relation to the data of other
persons (or situations).
Retain all the information attributable to the subject or situation in
question and create the categories of membership of the subjects (or
situations) in which subjects (or situations) are similar to each other.
Induction is a very important learning mode of the living creatures.
A simple example can illustrate how one learns by induction.
Assume that you have in front of you a person who has never seen the
containers in common use such as glasses, bottles, jars, cups, vases,
boxes, bottles, jugs, cups, tetra, and so on.
Without comment in succession you show real examples of objects
belonging to the categories described above.
The person may look, smell, touch and weigh the items shown.
After having examined a sufficient number of objects, the person will be
easily able to group objects into categories that include objects
similar to each other globally, focusing on some characteristics more than
others deemed irrelevant because it does not discriminate.
For learning occurred, I could submit another object in the form of
glass of different color, other materials and other weight still obtain
the location of the object in the category of glasses.
Also for the person in the induction training could make two cups of
categories: those without handle and with the handle (mug).
Learning has enabled the person to recognize the useful aspects of the
object to move from the particular to the universal and neglecting the
non-influential.
Know4Business realizes the acquisition of knowledge hidden in the data
and is able to explain the characteristics of each identified group.
Know4Business is applicable in any field (medicine, market research,
economics, fraud prevention, customer profiling, bank credit, market
segmentation, political polls, pharmacological research, security
systems, etc..).
Know4Business is able to determine an Index of Knowledge (Knowledge
Index) measurement of the knowledge contained in the input data
processed, allowing you to answer the question of "knowledge exists in
the archive?".
Know4Business makes available charts and statistics on the results.
Know4Business runs on Cloud Computing and will be available in a pay per
use on GAE (Google App Engine).

Visualizzazioni: 98

Allegati:

Contribuisci alla discussione

 


© 2011 Creato daa KNOW NET S.r.l. — P.IVA 02932470137.   Tecnologia

Badge  |  Segnala un problema  |  Termini del servizio

Attenzione!

I Forum sono stati creati per discutere argomenti di interesse comune tra i membri di Officine Italiane Innovazione.

Essi non sono appropriati per la diffusione di iniziative o proposte.

Per promuovere le Tue iniziative utilizza Bacheca Business Plus: solo così i Tuoi annunci verranno visualizzati in prima pagina e inoltrati a oltre 13.000 imprenditori.

Non ora