KnowledgeMiner (yX) for Excel  Overview

There are two different editions of KnowledgeMiner (yX) for Excel plus a time unlimited Free version
for evaluation purposes: The Silver edition for solving smaller and simpler modeling problems, and our Gold edition for
highperformance, highend professional knowledge mining needs.
Table 1: Key parameters of all versions.
Excel versions

Selforganized Analytical Model

Similar pattern search

Diagram
generation in Excel

Yearly fees

eBook

2011 2008

Forecasting, Modeling, Classification

Forecasting

Line, Prediction Interval, Scatter

No

Free


Table 2: Different key parameters.
Versions

Full support of multiple cores/CPUs

# of potential inputs supported

# of samples
supported

Model export to Excel

Unrestricted
modeling/
forecasting

Price

Free

No

5

500

Yes

No/No

Free

Silver

Yes

250

50,000

Yes

Yes/Yes

Show

Gold

Yes

350

Excel sheet

Yes

Yes/Yes

Show


Table 3: Common modeling technologies and their applicability
to different data set dimensions.
Sample Size n

Small
number of inputs m
(m < 50)

Medium
number of inputs m
(49 < m < 500)

Large
number of inputs m
(499 < m < 50K)

Very Small
(n < 30)

yX, [R]

yX

yX

Small
(29 < n < 200)

yX, [R, NN]

yX, [R]

yX

Medium
(199 < n < 10K)

yX, R, NN

yX, [R, NN]

yX

Large
(10K < n < 1M)

yX, R, NN

yX, [R, NN]

yX


• NN  Neural Networks and other Machine Learning methods.
• R  Statistical regression methods.
• yX  Selforganizing, highdimensional modeling in
KnowledgeMiner (yX) for Excel according to version, Copper, Silver, Gold, or on request.
• [.]  Can be applied under certain conditions only.

 Brings HighPerformance Personal Knowledge Mining to Excel users with unprecedented
ease of model building and ease of model deployment likewise.
 Hides the complex processes of knowledge extraction, model development, dimension reduction,
variables selection, noise filtering to avoid overfitted models, and model validation to the user.
 Uses the familiar interface and functionality of Microsoft Excel for storing and organizing the data,
building models, and working with the generated models.
 Selforganizes linear or nonlinear, static or dynamic regression models
along with a equation that describes the data from up to 50,000 input variables.
 Highdimensional modeling by implementing unique model validation methods
which significantly help answering one key question
if the generated model reflects a causal relationship and to which extent
or if it just models noise. This leads to the concept of Descriptive Power of a model.
 Generated analytical models can be used for StatusQuo or WhatIf prediction,
analysis, simulation, or optimization problems in Excel.
 Nonparametric forecasting models for fuzzy objects
based on Similar Patterns technology.
 Free PDF copy of the book by Mueller/Lemke "SelfOrganising
Data Mining."
How KnowledgeMiner (yX) is Different 
Starting from the basic ideas in cybernetics to
the important influential works of A. G. Ivakhnenko in selforganizing modeling,
KnowledgeMiner (yX) for Excel is based on a proven 50year history and development
of essential concepts in mathematics, statistics, information and computer science,
and systems theory. It implements 64bit multithreaded, selforganizing modeling
algorithms to make efficient use of multicore hardware and to significantly reduce
timetodeploy of models by the factor of 10 or more compared to other data mining tools.
KnowledgeMiner (yX) was designed to generate models objectively
by taking as much as possible work and subjective decisions from
the user to enable true knowledge extraction from noisy data, which
must only be minimally influenced by any expectations or assumptions
of the user respectively modeler. A consequence of this strict approach is simplicity and
consistent ease of use of the tool.
KnowledgeMiner (yX) for Excel is forecasting software, where you
can analyze and forecast processes that take place over time,
statistical analysis and modeling software to better understand
how a complex problem can be described
mathematically, and classification software in one tool.
It provides a fast and unique way for scientists and business users
to generate and apply reliable models from low to highdimensional
noisy data for a wide range of problems.
