Sports news subject term extraction method based on BiLSTM-CRF

• Main Authors: JIANG YIQI, ZHAO TONGZHOU
• Format: Patent
• Language: Chinese; English
• Published: 28.02.2020
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

The invention discloses a sports news subject term extraction method based on BiLSTM-CRF, and the method comprises the following steps: obtaining sports news from a website as training data, and extracting a title and a text of the training data; extracting topic sentences from the text, and extracting topic sentences of the training data; establishing a training set and a test set, dividing partsof the obtained titles and topic sentences into the training set, and dividing the rest parts into the test set; establishing a BiLSTM-CRF model, training the BiLSTM-CRF model by taking titles and topic sentences in the training set as objects, and extracting topic words of training data in the training set to obtain an optimal prediction model; and extracting titles and topic sentences of the sports news of which the topic words need to be extracted, and substituting the titles and the topic sentences into the optimal prediction model to obtain the topic words of the sports news of which thetopic words need to be ex