Document type identifying method and document type identifying apparatus

A computer-readable storage medium that stores a document type identifying program for identifying a document type of a document having word information on word strings and ruled line information on ruled lines written thereon, the document type identifying program causing a computer to execute a process comprising:

• receiving inputs regarding the word information on the word strings and the ruled line information on the ruled lines written on the document;
• storing a plurality of keywords used as keys that identify document types in a storage unit in association with each document type;
• generating, for each keyword, a plurality of word-sets, which are new combinations of a plurality of words obtained by decomposing keywords consisted of a plurality of words into each word composing the keywords and further extracting a plurality of words from the each word, to be checked for matching with the word strings written on the document by using the keywords stored in the storage unit at the storing;
• generating, based on the received inputs regarding the word information and the ruled line information at the receiving, grouped word strings by grouping the words written on the document and overlapping each other in terms of height when seen from a lateral direction, and aligning each of the generated grouped word strings;
• checking matching of each of the grouped word strings aligned at the aligning with each of the word-sets generated at the generating, obtaining, for each of the keywords, number of matched words with a highest matching rate between each of the grouped word strings and each of the word-sets, and calculating an evaluation value for each document type by using each of the obtained number of matched words, each evaluation value evaluating a possibility of each document type to be the type of the document;
• and
• determining, based on each evaluation value calculated for each document type at the calculating, the document type of the document having the word information and the ruled line information written thereon.

A document type identifying apparatus that identifies a document type of a document having word information on word strings and ruled line information on ruled lines written thereon, the document type identifying apparatus comprising:

• an input unit that receives inputs regarding the word information on the word strings and the ruled line information on the ruled lines written on the document;
• a document type identifying information storing unit that stores therein a plurality of keywords used as keys that identify document types in association with each document type;
• a word-set generating unit that generates a plurality of word-sets, which are new combinations of a plurality of words obtained by decomposing keywords consisted of a plurality of words into each word composing the keywords and further extracting a plurality of words from the each word, for each keyword by using the keywords stored in the document type identifying information storing unit, the word-sets being to be checked for matching with the word strings written on the document;
• a word string aligning unit that, based on the word information and the ruled line information received by the input unit, generates grouped word strings by grouping the words written on the document and overlapping each other in terms of height when seen from a lateral direction, and aligns each of the generated grouped word strings;
• a word-set matching unit that checks matching of each of the grouped word strings aligned by the word string aligning unit with each of the word-sets generated by the word-set generating unit, obtains, for each of the keywords, number of matched words with a highest matching rate between each of the grouped word strings and each of the word-sets, and calculates an evaluation value for each document type by using each of the obtained number of matched words, each evaluation value evaluating a possibility of each document type to be the type of the document;
• and
• a document type determining unit that, based on each evaluation value calculated for each document type by the word-set matching unit, determines the document type of the document having the word information and the ruled line information written thereon.