Apoala Project

.

Project Integration
Over the past year the Apoala Project has explored and developed of a number of information management and exploratory analysis techniques that may all be grouped under the rubric of "information science". The goal of this project is to integrate these disparate areas of research in a system that allows a domain expert to move seamlessly through his or her data in search of patterns and connections that would not otherwise be apparent.

Data Mining

Data mining involves the automated extraction of hidden information from databases. In our work, we are using a public domain software package called AutoClass to provide an unsupervised Bayesian classification of the readings taken at weather stations in the United States. Below you can see a data cube showing how a subset of the data has been classified. The main loaded variables in this classification are daily minimum and maximum temperature readings (represented here as the TMIN and TMAX axes). The four "point clouds" in the cube are four different classes that the data have been classified into. AutoClass performs fuzzy classification, and the shading of the symbols represents the likelihood of the data falling within its given class (only the most probable classes are mapped here).

..

.

Copyright © 1998-1999 Pennsylvania State University


.

	Project Integration Over the past year the Apoala Project has explored and developed of a number of information management and exploratory analysis techniques that may all be grouped under the rubric of "information science". The goal of this project is to integrate these disparate areas of research in a system that allows a domain expert to move seamlessly through his or her data in search of patterns and connections that would not otherwise be apparent. Data Mining Data mining involves the automated extraction of hidden information from databases. In our work, we are using a public domain software package called AutoClass to provide an unsupervised Bayesian classification of the readings taken at weather stations in the United States. Below you can see a data cube showing how a subset of the data has been classified. The main loaded variables in this classification are daily minimum and maximum temperature readings (represented here as the TMIN and TMAX axes). The four "point clouds" in the cube are four different classes that the data have been classified into. AutoClass performs fuzzy classification, and the shading of the symbols represents the likelihood of the data falling within its given class (only the most probable classes are mapped here).
..

.	Copyright © 1998-1999 Pennsylvania State University