Research And Application Of Smart Home NLP Algorithm Based On Neural Network

With the continuous expansion of the market size of smart homes and the rapid development of natural language processing technology,the research value of natural language processing technology and the application prospects in the field of smart homes have begun to rise rapidly.In order to better study and apply natural language processing technology,we first need to study lexical analysis,whose core research content is Chinese word segmentation algorithm.The research content is the basis for subsequent syntax analysis and semantic analysis,and also a key step based on smart home applications.Based on the above situation,this article will take the following two questions(1)the selection and optimization of the Chinese word segmentation algorithm based on the rule dictionary and the Chinese word segmentation model based on the neural network(2)under the above research,the smart home is used as the background The flow application structure of the control instruction has completed the following work:(1)Completed the optimization of the traditional MMSEG word segmentation algorithm,and tested the Bi-LSTM + CRF word segmentation model under different hyperparameters.(2)Completed the optimized MMSEG algorithm and the filtering algorithm system based on the Bi-LSTM and CRF word segmentation algorithm,and tested it.(3)According to the situation in the field of smart home,the screening algorithm system is applied to the lexical analysis.And completed the process application of smart home control,built a dictionary,syntactic library and semantic library,according to the complete process design,reached the mapping relationship between input sentences and control instructions,and finally tested the overall algorithm process accuracy.Finally,based on the above test results,the optimized MMSEG word segmentation algorithm has a certain improvement in the three evaluation indicators compared with the traditional MMSEG word segmentation algorithm.Compared with the two algorithms,the screening algorithm has improved in the evaluation index.The algorithm solves the problem of optimizing dictionary unregistered words in MMSEG.This algorithm solves the problem that the Bi-LSTM + CRF word segmentation model requires a large training cost when adding professional vocabulary or new vocabulary.In the overall algorithm flow test,the accuracy rate shows a high level,indicating that the process has certain application value.