Research On Classification Algorithm Of Food Nutrition Popular Science Text Dat

At the present stage,people’s living standards and consumption concepts have undergone earth-shaking changes.While pursuing material enrichment,they also put forward higher requirements for diet,especially paying more attention to food nutrition.With the rapid development and wide use of Internet technology,the public can more easily find scientific knowledge about food nutrition on the Internet.However,diversified contents and forms of publicity also lead to uneven credibility of information.In addition,the public’s lack of knowledge of popular science information about food and nutrition makes it easy for them to obtain wrong information,thus affecting their health.In this context,the food nutrition popular science information monitoring system has been developed and put into use,realizing the effective monitoring of food nutrition popular science information.The core technology of this system is the food nutrition text classification algorithm,but with the explosive growth of food nutrition popular science information data,the support vector machine model adopted by the current system has obviously been unable to meet the practical needs of fast and accurate processing of massive data,improving the classification performance of the algorithm has become an urgent problem to be explored.Therefore,the classification algorithm of food nutrition science popularization information monitoring system is improved and optimized in this paper.The research is mainly divided into the following two aspects:1、The SVM method based on machine learning is improved and optimized,and a sparrow search optimization SVM classification method is proposed.Support vector machine algorithm has obvious advantages for high-dimensional problems such as text classification,while the classification accuracy of traditional support vector machine algorithm is closely related to kernel parameters and penalty parameters of kernel function.In order to further improve the classification performance of support vector machine model,this paper uses the newly proposed sparrow search algorithm to optimize kernel parameters and penalty parameters,and then constructs the text classification model of sparrow search optimization support vector machine model.The proposed method and the traditional support vector algorithm are verified by experiments respectively.The experimental results show that the accuracy rate,accuracy rate,recall rate and F1 value of Sparrow search optimization SVM model are 83.8%,84.7%,81.6% and 83.8%,respectively.All indexes are significantly improved compared with the classification performance of traditional SVM algorithm.2、The classification method based on deep learning is introduced into the field of food nutrition popular science information classification,and a classification method based on the fusion attention mechanism of Bert-Bi LSTM-Text CNN is proposed.In order to solve the problem of low classification accuracy caused by polysemy,the latest BERTS algorithm is used to represent the sample data in the text representation stage.In the feature extraction stage,Bi LSTM was used to extract global variables and Text CNN was used to extract local variables.At the same time,the attention mechanism layer is designed at the output end,and the softmax classifier is connected to complete the classification.Finally,the classification performance of the classification method based on BERT algorithm and Bi LSTM-Text CNN fusion Attention was verified,and the classification accuracy rate,accuracy rate,recall rate and F1 values of the algorithm were 90%,89%,88% and 89% respectively.By comparing the classification performance of the two proposed methods,it can be seen that the classification method based on the fusion attention mechanism of Bert-Bi LSTM-Text CNN is significantly superior to the Sparrow search optimization support vector machine model in classification accuracy,accuracy,recall and F1 value.Therefore,the classification method based on the fusion attention mechanism of Bert-Bi LSTM-Text CNN can be applied to the "Food nutrition science Popularization Information monitoring system",so as to improve the classification performance of the system and provide a better basis for expert decision-making.This study also provides some references for the application of text classification algorithm in the field of food nutrition classification.