Research On Recipe Search And Intelligent Recommendation Method Based On Web Crawler

In smart home,intelligent cooking technology is indispensable,and how to automatically and efficiently obtain recipe information has become one of the research hotspots of intelligent cooking.Therefore,this thesis proposes to automatically obtain recipes through web crawler,recommend recipes from the perspective of energy consumption and evaluation of recipe pictures,and conduct system integration research.The main research work of this thesis:（1）Recipe search based on web crawler and energy-saving recipe recommendationIt is proposed to use web crawlers to automatically obtain the recipe information on the Internet,and adopt the breadth first strategy to grab the recipes of common recipe websites.Jieba word segmentation is used to segment the recipe text,so as to obtain the key words related to fire time sequence.The recipe energy consumption is estimated through fire time sequence and electric energy fine-grained energy consumption model,so as to recommend energy-saving recipes to users.Four popular recipes are taken as examples to verify the above method.（2）Evaluation of recipe pictures based on deep learning model and recommendationThe web crawler is used to capture the recipe pictures on the Internet to make data set samples,and the recipe pictures are divided into poor,medium and very good categories according to the color standard.On Py Torch platform,five classic CNN models including Alex Net,Google Net,Res Net,Dense Net and Mobile Net are tested and trained.The experimental results show that the Mobile Net model has relatively high recognition rate and small overhead,and the test recognition rate is as high as 78.4%.For the pictures facing the actual recipe,the recipe corresponding to pictrues identified as good category evaluations are recommended object,and the corresponding webpage link information is recommended to the user.（3）System integration and designDesign the overall architecture of recipe search and recommendation system based on web crawler,and use Python+Django framework to develop the main functions of the system.Use IFLYTEK sound cloud platform speech recognition and speech synthesis for the system to provide the speech engine,integration of the second chapter to achieve web crawler as subsystem grab recipes from the Internet information,analysis recipes text to calculate recipes energy consumption and recommended energy-saving recipes,deploy training completed in the third chapter and export CNN model,provide the recipe image evaluation mechanism and recommend high evaluation of the recipe.Most of the system functions are realized in this thesis.The research work in this thesis can be applied to intelligent cooking and building energy conservation,which is worthy of long-term in-depth research.Figure [47] Table [15] Reference [73]...