|
|
 |
|
||||||||||||||||||||||||
| "AC2 tests all possible combinations of database fields to find the criteria that answer the question you put to your data best and ranks all relevant criteria along a decision tree." Byte Magazine |
Data Mining uncovers
the highly valuable decision-making information hidden in your data
and makes it available to build more accurate predictive models and
performance improvements. To get the greatest possible benefits from
Data Mining, one needs a precise methodology
including phases of data manipulation, information discovery and validation,
and knowledge application.
AC2 is a user-friendly and powerful system that supports
all aspects of Data Mining. It provides a comprehensive set of tools
to access, select, prepare, and manipulate data. AC2's intuitive
and powerful Object-Oriented language allows you to structure the data
and enrich this structure with your domain knowledge. AC2's
knowledge discovery engine - based on state-of-the-art inductive techniques
- builds predictive models automatically. These are displayed in the
form of decision trees. You can easily
test and validate the models and deploy them in areas such as segmentation,
classification, estimation, and prediction.
click on thumbnail to view a bigger screen copy.
AC2 is a comprehensive toolkit that allows you to perform
advanced Data Mining tasks and develop powerful decision support systems
step-by-step. Its intuitive and user-friendly interface makes it easy
for data owners and IT specialists alike to build and maintain their
own Data Mining Systems faster than ever before.
Accessing and selecting data.
AC2 can directly access the leading databases as well
as spreadsheets and other packages. The imported data model is converted
into a class/relation OO model. Once imported, data may be sampled,
filtered, or clustered.
Data preparation and manipulation.
You can enrich the data model by including your domain knowledge
or by merging data from several sources. Specialized graphical editors
allow you to create and modify new entities, relations and fields.
Information discovery and validation.
You select a question to be answered and let AC2 explore
your data. AC2 builds predictive models that take into
account the data structures and relations. A variety of validation
techniques, such as bootstrapping and cross validation, make it
easy to test and validate the reliability of the models extracted.
Knowledge application. Once tested
and validated, predictive rules and models can be immediately put
to work within AC2 for prediction and decision-making
tasks, or for performing interactive "what if" analysis using AC2's
simulation module.
click on thumbnail to view a bigger screen copy.
Marketing: customer satisfaction
profiles, market segmentation, classification, purchasing patterns,
customer profiles
Direct Marketing: market segment
profitability, response surveys, return on targeted campaign
Finance, Insurance: scoring, risk
assessment, fraud detection
Human Resources: applicant profiles,
hiring policies, career evaluation
Industry: quality assessment,
assembly line failures
Health: biomedical research, epidemiology,
clinical trials
AC2 offers
a variety of decision tree building algorithms. It features advanced
functions such as interactive tree editing and modification, tree
pruning, and total user control on tree building parameters and
constraints. AC2 includes
a wide range of methods for providing maximum reduction of noise
and missing values. AC2's
graphical OO modeling language supports multiple inheritance, dynamic
links, relations, functions, taxonomie, facets, etc. AC2's
integrated script language allows partial or total automation of
the different phases of the Data Mining process. AC2 supports
all types of numeric and alphabetic data including integer, real,
formula, boolean, symbol, text, date, etc. It also handles such
constraints as range, multiple values and default values. 
click on thumbnail to view a bigger screen copy.
AC2 Shell:
comprehensive toolkit with OO graphical language, script language,
specialized editors and customization module for building Data Mining
Systems. AC2 C/C++
Libraries: libraries of C/C++ calls for embedding your AC2
application into other applications. AC2 Runtime
Kit: distribute your AC2 end-user applications in your
organization. AC2 Data
Mining tools are multi-lingual (English, French). They are available
on PCs running Windows 3.1/NT/95 in 16 and 32 bits versions, and
the following Unix workstations running Motif: SUN, IBM RISC System/6000,
BULL DPX20, HP 9000 series 700 and DEC Alpha.
