KBSI’s Personal Data Prospector (PDP®) toolkit allows you to perform systematic knowledge analysis using state-of-the-art data mining techniques on legacy databases. An innovative graphical interface provides advanced user interaction, making it possible for non-data mining experts to successfully apply PDP® for decision support applications.
Overview
KBSI’s Personal Data Prospector (PDP®) is a software toolkit that facilitates the rapid deployment of decision support solutions using data mining and knowledge discovery techniques. The PDP® toolkit provides a systematic knowledge analysis capability via the state-of-the-art data mining of legacy databases. An innovative graphical interface provides advanced user interaction, making it possible for non-data mining experts to successfully apply PDP® for decision support applications.
Central to this capability is the PDP® toolkit’s use of Knowledge Analysis Networks in performing the Knowledge Discovery Process (KDP).
The PDP® toolkit allows analytical data-mining processes to be specified in terms of process diagrams that help to automate the knowledge discovery process. Users can construct these diagrams via the PDP® toolkit’s graphical user interface, and the tool provides a collection of “building blocks”–including data manipulation elements, data processing elements, numerical algorithmic components, and data visualization components– that can be pieced together as diagrams on screen to define and execute any of a host of analytical processes. These building blocks can be assembled in a variety of ways, allowing users to construct tailored KDPs that address their particular decision information needs. The KDP building blocks available in the PDP® toolkit provide fundamental knowledge discovery functionality.
Features
- Data access and extraction blocks enable users to communicate/retrieve/export data from single or multiple data sources and to store processed or partially processed data for future use. Both accessed and stored data supporte a wide variety of legacy data storage formats (e.g., text files, spreadsheets, relational databases, etc.).
- Data transformation blocks capture the common transformation processes–aggregation, filtering, smoothing, cleaning, and mathematical transforms–that are employed on data sets to make them amenable to data analysis or interpretation. Knowledge production/discovery blocks are collections of focused search elements that extract particular types of knowledge from the data. These blocks are algorithmic components that provide the mechanisms for specifying and carrying out previously classified data mining techniques; i.e., techniques classified according to the type of knowledge to be extracted.
- Knowledge production/discovery blocks are collections of focused search elements that extract particular types of knowledge from the data. These blocks are algorithmic components that provide the mechanisms for specifying and carrying out previously classified data mining techniques; i.e., techniques classified according to the type of knowledge to be extracted.
- Knowledge interpretation blocks identify certain behaviors or characteristics that exist in the data and which need to be interpreted in the context of the particular knowledge discovery problem situation. Based upon the outputs from the knowledge production elements, knowledge interpretation elements enable the interpretation of this knowledge in terms of the business goals of the domain.
The PDP® toolkit’s innovations in design and user interaction greatly enhance the ability of non-data mining experts to perform structured knowledge discovery and apply their knowledge discovery processes in developing decision support applications. The web-based version of the PDP® toolkit makes it even easier for users to work with the tool, providing distributed and secure access to the data mining functionality through a web browser.
Since it was first introduced, the PDP® technology has been applied in a wide variety of areas ranging from threat-detection for homeland defense to manufacturing, aviation logistics decision support, and clinical and business decision making in medical heath care systems.