2D Mapping of Large Quantities of Multi-variate Data

摘要

A new method for intelligent or ontent dependent retrieval of objects from among a large quantities of multi-variate data is devised and explained. The method is based on the combination of two different approaches. One is the multi-branching decision tree and the second is Kohonen neural network. The new method allows a retrieval of similar or identical objects from a number of N objects (N being in the order of 10~6 and more) in a number of comparisons proportional to log_9N. The method was developed in the connection with the question how to map millions of multi-dimensional objects like spectra, structures, time-series of process variables, multi-component analyses of food or pharmaceutical products, etc.?. In order to show how the proposed method works, a small example of 572 objects (8-component analyses of various olive oils) is described.