Finding frequent substructures in chemical compounds
Luc De Raedt, Hannu T. T. Toivonen, Ross D. King · Lirias · 1998
The discovery of the relationships between chemical structure and biological function is central to biologi-cal science and medicine. In this paper we apply data mining to the problem of predicting chemical carcino-genicity. This toxicology application was launched at IJCAI’97 as a research challenge for artificial intelli-gence. Our approach to the problem is descriptive rather than based on classification; the goal being to find common substructures and properties in chemi-cal compounds, and in this way to contribute to sci-entific insight. This approach contrasts with previous machine learning research on this problem, which has mainly concentrated on predicting the toxicity of un-known chemicals. Our contribution to the field of data mining is the ability to discover useful frequent pat-terns that are beyond the complexity of association rules or their known variants. This is vital to the prob-lem, which requires the discovery of patterns that are out of the reach of simple transformations to frequent itemsets. We present a knowledge discovery method for structured data, where patterns reflect the one-to-many and many-to-many relationships of several ta-bles. Background knowledge, represented in a uniform manner in some of the tables, has an essential role here, unlike in most data mining settings for the discovery of frequent patterns.