Machine learning and statistics: a matter of perspective
Sally Jo Cunningham · 1995
: Information has become an important commercial commodity--indeed, possibly the most important product of the future. While we have well-developed technologies to store data, the analysis to extract information is time-consuming and requires skilled human intervention. Machine learning algorithms augment statistical analysis by providing mechanisms that automate the information discovery process. These algorithms also tend to be more accessible to end-users and domain experts. The two analysis methods are converging, and the fields have much to offer each other. 1. Introduction With the increasing availability of computers comes an ever-expanding amount of scientific and commercial data stored in machine-readable form. In 1989, it was estimated that 5x10 6 computer databases existed in the world. Given that it is also estimated that the amount of information in the world doubles every 20 months, it is likely that both the size and number of databases have grown dramatically since th...