Recommendation Scheme Design Of Aviation Safety Information Based On Machine Learning

Civil aviation is a complex giant system,and the safe and normal operation of the system requires a large amount of information exchange.Obtaining accurate and timely safety information is very important for carrying out aviation safety management.However,in the face of a large amount of safety information from various channels and many platforms,on the one hand,civil aviation practitioners have difficulty finding the information they need,and on the other hand,it is often difficult for information providers to target the audience of the information.In order to obtain a new way of sharing aviation safety information and realize personalized recommendation of aviation safety information,this paper designs an information recommendation scheme based on a recommendation algorithm combining content filtering and tags.First,analyze the information needs of civil aviation practitioners and build an information demand model for recommended objects.Based on the analysis and determination of six types of civil aviation jobs,the information needs analysis is carried out from two aspects of questionnaire survey and job requirements,and then the user’s subscription and historical behaviors are analyzed to obtain a collection of individual user information preference tags.The tag weight is calculated by the tag probability,and the weight is dynamically updated with the time decay constant and the feedback behavior type,and a user information preference model composed of tag words and weights is constructed.Finally,the mean value of the information preference vectors of all users in the same category is calculated,and the information preference model of this category of users is constructed.Then,the aviation safety information is marked with feature labels to build a recommended information resource model.Through information analysis,the recommended information resources are divided into seven categories.Chinese word segmentation and keyword extraction were performed on the information title,and then the existing tags of the information were manually added to establish a keyword set representing the information,and a tag library of various aviation safety information was constructed.After each information resource tag is quantified by the TF-IDF method,an information feature tag model is constructed.Finally,design and verify the recommended scheme.A recommendation algorithm combining content filtering and tags is used to build an aviation safety information recommendation model for new users,old users,new resources and existing resources.By simulating user data and information resource data,using Python language programming to design and verify the recommended scheme.The accuracy rate and recall rate in Top-N recommendation are used to determine the value of N,and the information corresponding to the top N in the cosine value ranking is recommended to the user.Then the N value is used to determine the threshold of the recommended information to determine whether to recommend a new resource.After research,this article has come to the conclusion: improved the recommendation algorithm based on the combination of content filtering and tagging,the design of aviation safety information recommendation scheme is carried out.By constructing the user information preference model and aviation safety information feature model,the cosine similarity of the user demand vector and the information resource vector is calculated in the vector space model to filter the recommended information,and an information recommendation list with N=6 is obtained.This scheme can realize the active push of aviation safety information,and provides a new sharing mode for aviation safety information.