Method and apparatus for the classification of data

摘要

An apparatus and a method for creating a database wherein the data entry, such as a journal entry in accounting, is the canonical record can be implemented utilizing a microcomputer. A plurality of data entries to be classified are separate records each comprised of one or more items having associated quantities and an entry number serving as a pointer to the record. Each item contains information including at least an item number or label and a quantity. At least two item indicators are then generated for each of the item numbers. A mapping function is applied to each data entry to assign the item indicators for the item numbers based upon the associated quantities. The item indicators for the data entry are sorted into ascending numerical sequence and an n-dimension sparse matrix is selected where "n" is the number of items in the data entry. If the present combination of item indicators is new, a design record is created for the database based upon the sparse matrix and including the item indicators, the associated quantity sums, the total number of data entries summarized in the design record and a pointer (a chain of entry numbers) to the records of the data entry detail. The quantities for the present data entry are added to the quantity sums and the entry number is stored in the pointer chain. After all the data entries have been processed, a search routine can be utilized to review the various design records as desired.