摘要 |
PROBLEM TO BE SOLVED: To analyze data at a high speed with high accuracy.SOLUTION: A processor calculates a plurality of first hash values corresponding to values of data items, for each first hash function, by being given to each of a plurality of first hash functions, on each learning input data. The processor selects a specific first hash value out of the first hash values, for each first hash function, and gives the specific first hash value to a second hash function, to calculate a plurality of second hash values, on each learning input data. The processor aggregates the second hash values corresponding to the specific first hash values, to generate a learning feature vector indicating characteristics of a learning input data group.SELECTED DRAWING: Figure 1 |