| Since the beginning of the 21 st century,science and technology have achieved rapid development,and computer technology has deeply affected people.People’s daily life has become more and more dependent on computer technology.In the past four or five years,the storage price per byte of computer memory and external memory has been steadily declining,and the operation and calculation speed of various computing control units have gradually become faster.Experienced a new round of development opportunities and opportunities.In the process of planting agricultural products,agricultural practitioners often face various problems;when encountering problems,agricultural practitioners need to consult a large number of books or solve problems based on years of accumulated experience,which leads to problem solving.Inefficiency and other defects.In recent years,the data related to the agricultural field has exploded,and how to use these data to support the development of the agricultural industry has become an urgent problem to be solved.This paper designs and implements an intelligent question answering system for agriculture.The specific work is as follows:1.Due to the thousands of years of writing habit that there are no natural separators between Chinese characters and between words;when performing Chinese reading tasks,Chinese word segmentation must be performed first,and based on a complete Chinese sentence The context segments Chinese characters and recognizes Chinese words composed of characters and characters.This topic improves the existing bidirectional LSTM Chinese word segmentation model to improve the accuracy and optimizes,and assigns different threshold weights to the forward LSTM layer and the reverse LSTM layer in the bidirectional LSTM network,to better simulate the human brain’s response to Chinese sentences.word segmentation,effectively improving the accuracy of Chinese word segmentation.2.A machine intelligent reading comprehension model is constructed.The model first simulates the comprehension process of humans when reading articles,then captures the relationship between the article,question and answer,and finally encodes and decodes the above relationship based on a two-way attention mechanism.,so as to build an intelligent question answering model.Experiments show that the intelligent question answering model in this paper is effective and accurate.3.An intelligent question answering model for agricultural field is constructed.First,use web crawlers to crawl agriculture-related texts from professional and well-known websites such as China Plant Net,China Farmer Net,China Agriculture Net,China Agricultural Information Net,and Baidu Encyclopedia.Then,a series of steps such as data cleaning,Chinese word segmentation,machine comprehension and model building are performed on the obtained text,and a neural network model of agricultural knowledge question answering is constructed.4.Design and implement an intelligent question answering prototype system.Based on JSP and Tensor Flow,the whole system function is realized on the Intel workstation,and the system deployment and testing are completed.The test results show that the prototype system has a good performance in the EM index,and the system performance can also meet the real-time requirements. |
