Improved Intelligent Semantics Based Chinese Sentence Similarity Computing for Natural Language Processing in IoT
Document Type
Article
Publication Date
1-31-2021
Identifier/URL
40779375 (Pure)
Abstract
It is desired in the Internet of Things (IoT) networks to apply natural language processing (NLP) technology to complete the information exchange tasks such as text summary or text classification between IoT devices. To achieve higher precision for the NLP of Chinese sentences, in this paper, we propose to utilize the deep neural network (DNN) to compute the semantic similarity of Chinese sentences. The proposed DNN consists of the input layer, the semantic generation layer, the concat layer, the dropout layer, the hidden layer, and the output layer. We propose to train the intelligent semantic similarity calculator sequentially to extract the semantic feature and the context information feature. After the offline training, the resultant configured intelligent semantic similarity calculator could evaluate the semantic similarity of Chinese sentences. Furthermore, we provide numerical analysis to demonstrate the improved similarity calculation precision and the consistency of the calculation accuracy in different fields.
Repository Citation
Ye, J.,
Zhang, L.,
Lan, P.,
He, H.,
Yang, D.,
& Wu, Z.
(2021). Improved Intelligent Semantics Based Chinese Sentence Similarity Computing for Natural Language Processing in IoT. IoT as a Service: 6th EAI International Conference, 234-246.
https://corescholar.libraries.wright.edu/ee/188
DOI
10.1007/978-3-030-67514-1_19
