Method and system for structuring case text

The invention discloses a case text structuring method and system, and the method comprises the steps: obtaining an input text, mapping each word in the input text into a vector representation through a word embedding method, and generating an input text sequence; constructing an LSTM model, training parameters of the LSTM model through training data, inputting the input text sequence into the LSTM model, and generating a hidden state of a plurality of time steps; constructing an attention mechanism model Attention, generating an attention score of each hidden state according to the hidden states of a plurality of time steps, calculating an attention weight of each hidden state through a normalized exponential function, and calculating a weighted hidden state sequence according to the attention weight; and constructing a random field model CRF, and calculating conditional probability distribution of the tag sequence according to the weighted hidden state sequence. 本发明公开一种病例文本的结构化方法及系统，该方法包括：获取输入文本，通过词嵌入方法，将所述