Lao named entity recognition method based on combined neural network

The invention discloses a Lao named entity recognition method based on a combined neural network, and belongs to the field of small language recognition in natural language processing. First, coding Lao sentence sequences by using Bi-LSTM (bidirectional long short-term memory model), and outputting...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: ZHOU LANJIANG, ZHANG JIAN'AN, MAN ZHIBO, LI XUANDA
Format: Patent
Sprache:chi ; eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The invention discloses a Lao named entity recognition method based on a combined neural network, and belongs to the field of small language recognition in natural language processing. First, coding Lao sentence sequences by using Bi-LSTM (bidirectional long short-term memory model), and outputting character vectors; then, slicing and segmenting character vectors, and inputting character vectors into Bi-RNN (bidirectional recurrent neural network) model to obtain information representation vectors of internal units of fragments; on the basis, using Lookup operation to obtain an overall vectorrepresentation of the segment, then splicing the obtained segment information representation vector, the overall vector and other feature vectors as features, inputting the features into the neural network model, and performing Lao named entity recognition training. The recognition effect of the method is obviously superior to that of a traditional statistical learning method, and the recognitionperformance equivalent to