A M P dvanced

Transcription

A M P dvanced
AdvancedMiner Professional
a system for data transformations, exploration and Data Mining
AdvancedMiner Professional
a system for data transformations, exploration
and
Data Mining
AdvancedMiner Professional is an advanced software solution capable of carrying out complex
analytical projects of various kinds. Its effectiveness has been proven in numerous business
projects.
The possible applications of AdvancedMiner Professional include:
increasing the effectiveness of marketing campaigns,
lowering churn risk,
evaluating the reliability of borrowers (Credit Scoring),
creation of data marts for analyses or reporting,
automation of the process of verifying the quality of collected data.
Besides the AdvancedMiner Professional, StatConsulting also offers training and other support services
oriented toward increasing the effectiveness of working with the system.
Deployment variants
AM Environment – a basic work environment
for analysts,
AM Engine – a built-in efficient computational
database engine,
AM DataWorkshop – an application for multifaceted
work with data and reports,
AM Predictor – a solution dedicated to customer
behavior predicting and modeling projects,
AdvancedMiner Professional
AM ScoringCard
AM Modeler
AM
Predictor
AM
Clustering
AM Environment
I
AM
DataWorkshop
AM Engine
AM Clustering – a tool which enables effective
execution of data segmentation projects,
AM Modeler – a software package consisting of AM Predictor and AM Clustering,
AM ScoringCard – a tool for constructing and validating scoring cards as well as ad-hoc analyses
performed during Credit Scoring and Credit Rating projects.
System description
AdvancedMiner Professional is an integrated environment dedicated to the development of analytical
projects. The system offers various tools supporting the work of analysts and programmers. Our software
not only includes data processing tools, but also provides a wide range of statistical algorithms and
Data Mining methods, which can be used to construct effective analytical models.
AdvancedMiner Professional is designed for novice and advanced analysts alike. Beginner users may
build their first models and experiment with specific algorithm settings using the graphical interface and
default settings. Advanced analysts and experts will discover a valuable tool which can be customized
and extended according to their needs. This is possible thanks to the built-in scripting language Gython
(based on Python) which provides the functionality of a programming language augmented with special
functions for effective data analysis.
Additionally, our system supports group work: the analysts can share metadata objects through
common repositories or local and remote directories. Scripts and libraries created by the users can be
managed using the CVS version control system; it is also possible to share the processed data among
users. The system can work on the majority of popular operating systems and can be connected
to various database systems.
The graphical environment of AdvancedMiner Professional
AdvancedMiner Professional tools
Working with data sets
The system offered by StatConsulting can be successfully employed by companies which operate
on large data volumes on a regular basis (i.e. telecommunication, banking).
AdvancedMiner Professional can import data from:
practically any relational database supporting the ODBC/JDBC standard
(including MS SQL, MySQL, Oracle, Sybase),
data warehouses,
CSV files,
spreadsheets.
The user may not only explore but also edit databases, as well as create new tables using the built-in
scripting language. It is possible to use SQL queries and special language constructions for effective
data transformations.
Databases
AdvancedMiner Professional is equipped with pre-defined transformations (e.g. binarization,
standardization). These tools make the work of the analyst more efficient, shortening the time
used for data preparation. The users may define custom transformations as well.
Working with variables
AdvancedMiner Professional offers the user a wide range of functionality supporting the processes
of data exploration. The system comes equipped with standard statistical procedures as well as
a specialized interactive tool for graphical data exploration, which is capable of:
visualization of data distributions,
exploring variable correlation with the target,
creating virtual variables supporting the analysis,
grouping of variable values,
variable correlation analysis (e.g. Pearson, Correlation Ratio, ChiSquare).
Correlation matrix
Categorical variables
Modeling
AdvancedMiner Professional provides
tools for performing various tasks, such as:
classification,
approximation,
clustering,
association rules,
survival analysis.
The algorithms available for analytical
model construction include, for instance:
classification trees,
linear regression,
logistic regression,
k-means,
Classification tree
association rules.
AdvancedMiner Professional can be used to evaluate and compare the quality of models using tools
like LIFT, ROC, K-S or Confusion Matrix. The whole process can be done using an intuitive interactive
graphical user interface.
Model quality reports can be exported to MS Office or OpenOffice file formats.
Data Mining models quality assessment
Scoring code
The system can generate scoring code in various formats, offering the possibility of using the
constructed models independently from AdvancedMiner Professional.
This feature simplifies and speeds up the implementation of models through easy integration with
other IT systems in a company.
Reporting
With AdvancedMiner Professional it is possible to generate periodical reports containing a wide
range of statistics and comparisons. The reports are generated automatically or semi-automatically
and require little preparation time. AdvancedMiner Professional also supports custom scripts for
creating reports with elements like formatted text, autotext, calculation results, tables or charts.
Reports can be exported to applications like MS Office, OpenOffice and as PDF or HTML documents.
Part of a report generated in AdvancedMiner Professional
System architecture
AdvancedMiner Professional is based on a Client – Server architecture, where the server may be
physically distributed over several computational units. The security of communication between
the computational units is ensured by the SSL (Secure Socket Layer) protocol. This solution enables
secure multitasking work. While waiting for the outcome of the defined task, the analyst may start
preparing another task and perform it on the same or different server. This architecture speeds up the
process of data analysis and ensures full utilization of the available resources. With AdvancedMiner
Professional it is possible to set up an effective and scalable system for data transformation and
analysis by utilizing low cost PC type computers even with different operating systems.
IDE Client
IDE Client
SSL
SSL
SSL
Computation
Server
Computation
Server
Metadata Server
Load Balancer
Computation
Server
Metadata Server
Computation
Server
AdvancedMiner
Professional Server
SSL
GDBase – dedicated
analytical database
engine
Database
Database
The architecture of AdvancedMiner Professional
Compatibility
AdvancedMiner Professional is based on well-tested Java technologies, providing platform
independence. The system operates in MS Windows as well as in operating systems from the Unix
family (including Linux).
AdvancedMiner Professional is compatible with relational database management systems which
provide the JDBC/ODBC interface (e.g. MySQL, MS SQL, Oracle, Sybase SQL Anywhere Studio).
An internal database – GDBase - is also available for the users.
AdvancedMiner Professional advantages
Integration with external IT systems
importing data from various databases (MS SQL, MySQL, Oracle, Sybase and other supporting
the ODBC/JDBC standard), data warehouses, text files and spreadsheets,
iscoring code generation which may include models and data transformations,
ireporting integrated with MS Office and OpenOffice,
icompatibility with MS Windows and the Unix family of operating systems.
Distribution of computations
speed up of the data analysis process by setting up the server to be physically distributed over
several computational units,
full utilization of the available resources by building up an effective, scalable and low cost
system made up of PC type computers with different operating systems (e.g. Linux, MS Windows).
Data processing
no restrictions on the number of columns in the processed tables,
high efficiency of operations handled by the integrated analytical database engine (GDBase),
capability of working with very large databases (such as 5 billion records) on a dedicated server
as well as PC-class workstations.
Integrated modeling environment
exhaustive data exploration functionality accessible through an advanced interactive graphical
user interface,
extensive possibilities of comparing models during prototyping and testing,
easy construction of hybrid models combing different analytical methods.
AdvancedMiner Professional advantages
Wide range of applications
such as customer data transformations and exploration, Credit Scoring, Churn analysis, analysis
supporting Cross/Up Selling, customers segmentation, LTV and Survival analysis.
Effectiveness
system flexibility due to the built-in scripting language designed for working with data
and models,
customization of the system to meet the individual needs of the client, as well as construction
of dedicated solutions by StatConsulting,
an advanced editor with model building support and a wide range of functionality not available
in other tools of this type,
uniform object oriented approach facilitating the effective use of different analytical methods,
group work support including CVS (Concurrent Version System), metadata and database sharing.
If you wish to obtain more detailed information about AdvancedMiner Professional,
feel free to contact us. We will gladly provide you with further details
and answer your questions.
About StatConsulting
StatConsulting was established in 2001 in Warsaw, Poland. We offer analytical support for risk
management (credit, market, operational risk with regard to the New Basel Capital Accord Basel II and Solvency II) and in the field of customer behavior modeling (Customer Intelligence,
analytical CRM).
We use our expert knowledge to conduct analytical projects for companies operating in highly
competitive business sectors. StatConsulting specializes in providing analytical services for the
financial sector and telecommunication companies. We completed numerous analytical projects
involving the construction of different types of models (incl. scoring and analytical CRM). We have
broad experience in building, validating and monitoring data analysis models including Application
Scoring and Behavioral Scoring. This also includes the building of analytical repository, data
preparation, integration with data from external sources, model building and implementation.
The high quality of the services provided and our experience in conducting and implementing
analytical projects enabled us to reach a significant position among other companies offering
data analysis.
StatConsulting clients
Banking and finance
AIG Bank S.A.
Alior Bank S.A.
Bank BPH S.A.
Bank Pocztowy S.A.
Bank Ochrony Środowiska S.A.
BRE Bank Hipoteczny S.A.
BRE Leasing Sp. z o.o.
Euro Bank S.A.
Europejski Fundusz Leasingowy S.A.
(major leasing company in Poland,
Crédit Agricole financial group)
Kruk S.A.
Mercedes-Benz Bank Polska S.A.
Mercedes-Benz Leasing Polska Sp. z. o. o.
PolCard S.A. (card payment services)
PTE PZU S.A.
PZU Życie S.A.
Raiffeisen-Leasing Polska S.A.
Toyota Bank Polska S.A.
Telecommunication
Polkomtel S.A.
(a major Polish GSM mobile operator)
Pharmaceuticals
Cegedim Polska Sp. z o.o.
Eli Lilly Polska Sp. z o.o.
Retail sales and FMCG
Bertelsmann Media Sp. z o.o.
Carrefour Sp. z o.o.
MDS Poland Sp. z o.o.
Industry
KGHM Polska Miedź S.A.
Klingspor Sp. z o.o.
Statoil Polska Sp. z o.o
Direct marketing and market research
ASM – Centrum Badań
i Analiz Rynku Sp. z o.o.
Premium Club Sp. z o.o.
StatConsulting Sp. z o.o.
ul. Wołodyjowskiego 38a
02-724 Warsaw, Poland
tel.: (+48) 22 / 847 97 17
fax: (+48) 22 / 499 45 31
e-mail: [email protected]
www.statconsulting.eu
European Funds for the development of innovative economy
Project co-financed by the European Regional Development Fund