A Chinese text classification model based on radicals and character distinctions

Chinese characters are generally correlated with their semantic meanings, and the structure of radicals, in particular, can be a clear indication how related to each other. In simplification movement, some different traditional have been transferred into one simplified character (many-to-one mapping), resulting phenomenon ’one corresponding many characters. Compared characters, contain richer structural information, which is also more meaningful understanding. Traditional approaches text modelling often overlook content role human cognitive behaviour process comprehension. Hence, we propose classification model derived from construction methods evolution The consists two branches: traditional, an attention module based on radical branch. Specifically, first develop sequential obtain sequence information texts. Afterwards, associated word using part head as medium designed filter out keywords high differentiation among auxiliary units. An then implemented balance importance keyword particular context. Our proposed method conducted three datasets demonstrate validity plausibility.