Word-splitting estimation model learning device, word-splitting device, method, and program

摘要

To provide a word division estimation model learning device, a word division device, method and program capable of learning a model for accurately dividing a text into words.SOLUTION: A word division estimation model learning device is configured so that: an input part 10 receives a text to which a word division label is applied, the word division label indicating whether or not there is a word division boundary for every character in a text; an existing analysis tool division result acquisition part 30 acquires a result obtained by dividing the text into a prescribed character string unit; a word division model learning part 32 learns a word division model 40 which is a neural network in which, a word division label applied to each character in the text is an output, with each character in the text and a result obtained by dividing the text into a prescribed character string unit as inputs, based on the word division label applied for every character in the text, and a result obtained by dividing the acquired text into a prescribed character string unit.SELECTED DRAWING: Figure 1