STATISTICA QC Miner


          

Contents

Summary

 STATISTICA QC Miner is a powerful software solution designed to monitor processes and not only identify but also ANTICIPATE problems related to quality control and improvement with unmatched sensitivity and effectiveness. The products combine the most powerful tools for QC and SPC with cutting-edge data mining technology.

This specialized version of STATISTICA Data Miner is offered either (a) as a complete "turn-key" (deployed) solution, custom tailored by StatSoft engineers to your specific needs, and/or (b) as a comprehensive set of tools that enables you to easily build new or customize existing solutions. STATISTICA QC Miner integrates the complete functionality of STATISTICA software for quality control and improvement with STATISTICA Data Miner software for uncovering hidden trends, explaining known patterns, and predicting the future:

  • All quality control charts, process capability analyses, experimental design procedures, or Six Sigma methods and charts are integrated with a comprehensive library of cutting-edge techniques for exploratory and predictive data mining.
  • QC Miner projects can be connected to live data streams and automatically updated as new data (samples, observations) become available.
  • QC Miner projects may also actively poll external databases to retrieve samples at predefined intervals.
  • QC Miner can process data sets of practically unlimited size; for example, you can screen even millions of characteristics.
Customized Solution Packages To Match Your Specific Needs

Several specialized Solution Packages are available for users of STATISTICA QC Miner:

  • General Quality Process Control Applications Package

  • Yield Management Applications Package for the semiconductor industry

  • Root Cause Analysis Application Package
    		•

    Go Beyond Simple Quality Control:
    Don't Just React to Problems, ANTICIPATE Them!
    • Predict quality problems before they happen, using neural networks, regression trees, generalized linear models, or other procedures from the vast library of cutting-edge statistical and machine learning methods.
    • Continuously monitor and improve expected (predicted) Return On Investment (ROI); predict process yield in real time, at each step of complex (manufacturing, service delivery, etc.) processes, and use automated machine learning algorithms to generate recommendations for modifying settings of tools, machines, etc.
    • Discover root causes of quality or process problems, applying the most advanced methods for feature (variable, factor) selection available.
    • Customize the system to monitor your (manufacturing, service delivery, etc.) processes in real time, learn from mistakes, and programmatically generate suggestions for improvement.
    • Incorporate active, real-time, on-line data mining methods in your data-driven Six-Sigma program: Provide not only instant feedback, but automatically identify the variables and factors that affect your processes.
    • Do not simply react to the results of your processes: Anticipate yield, quality, and areas for improvement.
    • Via STATISTICA's Client-Server (Web browser-based) installation, you can generate and deploy solutions for process improvement from anywhere in the world via the Internet, and communicate solutions to clients or other stakeholders worldwide.

    How it Works

     STATISTICA QC Miner consists of the complete set of tools available via STATISTICA Data Miner. In addition, the complete functionality of STATISTICA Industrial Solutions and Six Sigma Tools is available and are integrated into the data mining projects:

    • Multiple data (process) streams (or data sets) can be connected into a single project so that users can simultaneously compare, monitor, and analyze multiple processes, machines, parts, etc.
    • Multiple projects can process and analyze the same data streams.
    • Multiple data streams can be "routed" through the same nodes, applying the same engineering specs, quality control limits, sampling methods, analyses, etc. to all data streams. Users set up charts, analyses, or data filtering methods only once, connect new data or variables as they become available, and then connect multiple data streams to the same neural network monitor.
    • QC Miner projects can be customized using the syntax of industry standard languages (VB, C++). For example, users can create templates for different processes, variables, or applications.
    • QC Miner projects are fully programmable. This enables users to automatically develop highly customized solutions integrated with the local environment; for example, retrieve random samples from one or multiple remote database servers, clean the data, and update all charts, analyses, and learning algorithms.
    • Powerful Client-Server version (that supports distributed processing and scales to multiple servers) is also available.
    • A Web-based version, with browser-user interface is also available.

    Customize all Projects using Industry
    Standard Programming Languages

     STATISTICA QC Miner projects are created from analysis, data filtering/cleaning, or other nodes that encapsulate the respective analyses or data management operations. These nodes are "thin wrappers" written in the industry-standard open-architecture Visual Basic language, which call and execute the (extremely fast and efficient) STATISTICA analytic and data management routines. By using the industry-standard and convenient Visual Basic language as the "glue" to connect the process streams through the project, it is extremely easy to customize the system by incorporating custom routines written in Visual Basic, Java, C, C++, C#, or by calling programs supplied by other vendors.
    • There is no esoteric (specialized or STATISTICA-only) language that you need to learn; a huge pool of skilled and trained programmers already exists, probably within your team!
    • You do not need to commit to a single vendor or set of consultants to set up and maintain your projects; you determine how to modify projects, or expand the system.
    • Incorporate custom transformations, error handling, alarm notification, or analyses into your custom projects. For example, a neural network can send a warning when the probability of quality problems within the next 1 hour exceeds 50%.

    In-place Database Processing

     Unlike traditional industrial statistics and quality control software, STATISTICA QC Miner can be flexibly connected to data residing on your local computer, or to data in extremely large remote databases for in-place database processing, without requiring that a copy of the data be created on your local machine. Regardless of where and how the data are stored, or how large the databases containing process and quality-relevant information may be, STATISTICA QC Miner will easily and seamlessly connect to and integrate with your existing systems.

    There is practically no limit to the size of your data stream; STATISTICA QC Miner can screen unlimited numbers of characteristics (i.e., variables or features) and records.

    Using STATISTICA QC Miner
    as a Silent Monitor and Learner

    Like all procedures in the STATISTICA family of products, the procedures in STATISTICA QC Miner can be accessed from other applications (that support the industry standard Component Object Model or COM interface), and run "silently" inside the other application so that no actual STATISTICA QC Miner user interface will ever be displayed. For example, your existing quality and process control monitoring applications may incorporate calls to STATISTICA QC Miner to access the tremendous power of cutting-edge data mining, machine learning, and statistical algorithms.