Chapter 10. Introduction to AdvancedMiner

Table of Contents

Introduction
AdvancedMiner Client
AdvancedMiner Server

Introduction

AdvancedMiner is an integrated analytical environment developed by Algolytics. It is a software suite designed for extracting business knowledge from large data sets. AdvancedMiner supports the complete cycle of the Data Mining process:

  • data import
    • importing CSV files
    • access to multiple databases
  • data transformations
    • typical transformations
    • scripting language with access to data and SQL support
  • data exploration
    • statistical procedures and tests
    • visualization tools
  • model building
    • classical tasks: classification, approximation, segmentation
    • other tasks: survival modeling, time series, association modeling
  • model testing
    • access to model statistics and model visualization
    • model quality assessment and quality models comparisons
  • model application
    • using models in scoring processes
    • exporting models in the form of scoring code
  • reporting and monitoring
    • reprting of data transformatiosn and modeling processes
    • building customized reports using scripts and specialized APIs

Each step is fully supported by AdvancedMiner. Moreover, the steps are highly integrated with each other.

AdvancedMiner is based on the CWM Metadata concept. In AdvancedMiner the term metadata is used to refer not only to information about data, but all the information required to describe the Data Mining process, including modeling tasks, transformations, models and model statistics.

AdvancedMiner is a client-server application. It consists of two separate pieces of software and a separate modules package:

  • AdvancedMiner Client - the graphical user interface for the application and analytical tools.
  • AdvancedMiner Server - the server application which does the computations and stores the metadata objects
AdvancedMiner has its own database called GDBase. This database is optimized for analytical tasks. It is highly recommended to use this database, especially for large data tables.

AdvancedMiner IDE is an application that performs computations on a distributed server. Thanks to the platform's architecture, the computations can be executed on several computers simultaneously and independently of the physical distance between them. The user has full control over all the computations.

Certain processes (which do not refer to metadata objects) are executed locally on the client and other are run on the server. The user can track all the processes. However the processes of one user on the server are invisible to other users. It is also possible to terminate processes running locally or on the server.

AdvancedMiner IDE is designed to support group work. The platforms provides tools for:

  • metadata sharing – metadata can be shared on the AdvancedMiner Server in the form of password-protected repositories,
  • script sharing – scripts are stored as text files, so it is possible to use version control systems (CVS client is included in the system),
  • data sharing – a remote database can be simultaneously used by many users.