Artificial Intelligence
Investigating the effects of color
Communications of the ACM
| Hi-index | 0.00 |
Computer programs for graphing and analyzing time series data are widely available. For large data analysis applications, however, the analyst may invest a great deal of time navigating an ocean of data in order to find the relevant and interesting pieces. By making this process of discovery easier we can improve the productivity of the analyst. In this paper we describe a data analysis system composed of an eclectic combination of pattern recognition, artificial intelligence, and digital signal processing with the goal of providing some of the right tools. The machine is used to accept abstract descriptions of interesting or anomalous data and then to bring that data quickly into the user interface. The same tools can screen large datasets in the analyst's absence. The human analyst spends less time wading through graphs and numbers and more time answering the question of the day. Our goal is to empower the analyst by providing a higher-level language with which to manipulate, visualize, and restructure the semantic concepts of the domain.