| 摘要 |
PROBLEM TO BE SOLVED: To provide a k-nearest neighbors associative memory that effectively realizes a k-nearest neighbors algorithm.SOLUTION: A k-nearest neighbors associative memory 100 includes: a clock counting type associative memory 10 that holds R pieces of reference data and outputs, for each of the R pieces of reference data, a match signal that becomes active when a clock count corresponding to a distance between the reference data and given search data has been reached; and a k-nearest neighbors clustering circuit 30 that, each time at least one of the R match signals output from the clock counting type associative memory becomes active, selects a piece of class data, out of R pieces of class data representing classes of the R pieces of reference data, corresponding to each of the at least one active match signal, until k match signals out of the R match signals become active, and determines a class having a largest number of pieces of data when the selected total k pieces of class data are classified by classes.SELECTED DRAWING: Figure 1 |
