Machine learning from agricultural databases: practice and experience
Stephen Garner, Geoffrey Holmes, Robert John McQueen, Ian H. Witten · 1996
Per capita, New Zealand is reputed to be one of the world's leading collectors of information in databases. The country, quite rightly, places a substantial investment in stored knowledge. As databases grow, however, interest tends to shift from techniques for retrieving individual items to large-scale analysis of the data as a whole. Data can be analysed to view trends, pick out anomalies, discover relationships, check that policy is turned into practice, and so on. With large databases, such analyses can be costly. At Waikato University we are examining these issues in the context of the agricultural sector of the economy. Our project centers around techniques of "data mining" or "machine learning," which automatically analyse large bodies of data to discover hidden dependencies. Other methods of exploratory data analysis---many of them interactive---are being investigated as an adjunct to the machine learning algorithms. This paper reviews the software that has been developed to sup...