Method for expanding LIWC vocabulary

The invention provides a method for expanding a LIWC vocabulary. The method includes steps S1 of applying an attention mechanism to construct an i-th layer context representation of a target word according to a semantic element of the target word and an i-1 layer label of the target word; step S2 of inputting the i-1 layer label, the i-th layer context representation and an i-1 layer hidden stateof a recurrent neural network, to acquire an i layer hidden state; step S3 of applying the i layer hidden state to acquire an i layer label of the target word; step S4 of repeating steps S1 to S3, andobtaining each layer of label of the target word layer by layer; wherein m>=i>=1, i is an integer, and m is the total layer number of the target word. According to the method for expanding the LIWC vocabulary, the LIWC vocabulary is automatically expanded based on the attention mechanism combined with the recurrent neural network of the semantic element, errors caused by artificial expansion areavoided, the accuracy of the LIWC vocabulary annotation is improved, and the labor costs and time costs are reduced.