Research On Map Recommendation User Model Based On Text Processing

Earthquakes are one of the most harmful natural disasters to human survival.Earthquake emergency is to timely obtain disaster information and make relevant measures during the earthquake disaster emergency period to reduce earthquake casualties and property losses.In the earthquake emergency period,how to make effective and effective decisions quickly and minimize the damage and loss of earthquake disasters is the key to earthquake emergency.Establish a user model of earthquake emergency,organize spatial information according to the specific needs of different types of users,and quickly transmit spatial information that meets the needs of current users,so as to provide spatial information services for different users in earthquake disasters more accurately and effectively,so that they can better Efficiently complete tasks in earthquake emergency.The user model of the earthquake emergency thematic map is constructed for the production and recommendation of the earthquake emergency thematic map.The earthquake emergency map is a map for the four types of users,including command users,rescue users,disaster-affected users,and the general public,to understand the disaster situation,command and rescue,and disaster analysis and evaluation.Due to the different needs of various users for seismic information,the general earthquake emergency thematic map cannot often meet the needs of many types of users.Therefore,we need to establish a user recommendation system based on the earthquake disaster emergency thematic map,which can recommend the thematic map of the most concerned elements for specific users at the first time of the earthquake disaster.In the user recommendation system of the earthquake disaster emergency thematic map,user modeling is the core technology of the recommendation system.This study is based on text information processing,combined with user category information of earthquake disasters and demand information in emergency process,etc.The user modeling research of the graph mainly includes the following three aspects:(1)The earthquake disaster text is processed.The accurate extraction of user information is the key to ensuring the accuracy of user model information expression and improving the recommendation performance of emergency thematic map users..In this paper,the text data in the earthquake theme webpage is collected in a targeted manner,which avoids data redundancy and waste of network resources.Text classification and clustering are carried out by user classification information,time information and location information in the text,which is obtained with user interest.Thematic and purposeful text sets,through the establishment of the earthquake theme user dictionary,the text segmentation and feature word extraction processing,Finally,the TF-IDF algorithm is used to calculate the seismic related feature word frequency,and the user interest model is established.(2)Establish a user model for earthquake emergency map recommendations.Establish user multi-dimensional information such as the type,time and location of earthquake emergency map users,construct a user dictionary containing scene feature words,emergency demand feature words and special feature words,and analyze the shortcomings of VSM in expressing user information of earthquake emergency maps.A representation method of a multi-dimensional vector space user model based on user,time,and interest.(3)The validity and accuracy of the multidimensional vector space user model are verified by experiments.According to the verification data analysis,the recommended results obtained by using the multidimensional vector space model are consistent with the user's perception of disaster emergency.Through the Perplexity model,it is verified that the multidimensional vector space model is superior to the traditional vector space model in terms of text prediction ability and generation ability;the results obtained by the model are finally recommended by the system.