 |
 |
 |
 |
 |
 |
 |
 |

 |
 |
 |
 |
 |
|
MERKUR MINER PLUS - Technology
Architecture
MERKUR MINER PLUS is a Windows desktop application written in C++ and can run on all 32-bit Windows platforms.
The technological core of MERKUR MINER PLUS is based on hybrid OLAP technology combined with a data mining library. On the one hand, hybrid OLAP technology makes it possible to process extremely large volumes of data (several million datasets) with MERKUR MINER PLUS and on the other hand it simultaneously supports many dimensions (several hundred). Conventional OLAP servers often support far fewer than 100 dimensions and can thus not be used for many datasets.
While OLAP technology allows the user to analyse and visualise historical datasets, the integrated data mining library can be used for forecasts and classifications as well as for creating scoring models.
Data import
Data import in MERKUR MINER PLUS takes place via ODBC or OLE DB. Smooth access to the data of all common databases such as Oracle, Sybase, MSQL Server and Informix is guaranteed via these two database interfaces.
Analytical functions
A comprehensive palette of analysis objects is available to visualise and navigate through data. The spectrum ranges from pivot and contingency tables to a variety of diagram types right down to geographical maps. All of these analysis objects are equally suited to drill down, roll up and the selection of data.
The data mining object classifies data and can thus be used to create responders, cancellers, new customers and scoring models. When it comes to data mining technology, decision trees are used, providing the advantage of interpretable results. By integrating data mining into the OLAP architecture, it is possible to use data mining to identify interesting data segments (e.g. lucrative customer groups) and then to visualise and analyse these with the various analysis objects.
The data mining models created can be exported into PMML (Predictive Model Markup Language, an XML standard for the exchange of data mining models). It is also possible to use MERKUR MINER to import PMML models of other data mining systems (e.g. prudsys DISCOVERER).
Results and export
MERKUR MINER PLUS is equipped with a powerful report generator. All analysis objects can be saved in one report. This enables the user to define his own chapter structure and to assign different analysis objects to each chapter or section. Explanatory text can also be added to each object. The output of the report is in RTF format which is supported by all common office systems , making it easy to rework reports after they are created.
The analysis automation function of MERKUR MINER PLUS makes it possible to update data and reports on a daily, weekly or monthly basis. These can then be printed out or sent by email.
|
|
|
 |
|
 |
|
|