Text Recommendation Based On Abstraction Level Of Concept

With the rapid development of the Internet,the scale of text data is getting larger and larger,which increases the cost of information acquisition while bringing rich resources to users.Traditional text recommendation systems make recommendations mainly based on the relevance of the text to the semantic or topical needs of users.However,because users have different levels of knowledge or specific purposes,they may require the degree of abstraction and difficulty of the recommended text while requiring semantic or topic relevance.As a result,current text recommendation technologies cannot meet the requirements of users.The following two aspects are completed in this paper:(1)First,by combining cognitive linguistic knowledge and mining text reading methods hidden in user behaviors,this paper analyzes the user's current level of abstraction cognition,and then combines the abstraction of concepts in text with the cognition in cognitive linguistics to recommend text.Besides,users can also choose the abstract and specific directions to adjust the recommended content.(2)Second,to further recommend learning content,this paper analyzes the internal associations and laws between text abstraction and text difficulty in physical,information,and social spaces,and analyzes the user's learning behavior in the law of change in abstraction based on the knowledge of cognitive linguistics,which extends the features of text recommendation algorithms in different spaces and improves the efficiency of text recommendation.This paper uses Wikipedia text as the data source to experimentally verify the above method.Experimental results show that abstraction-based text recommendation can meet users' needs for text content with different levels of abstraction.At the same time,in the physical,information,and social spaces,we draw the following conclusions:(1)In physical space and information space,as the text difficulty increases,the overall level of text abstraction first decreases and then increases.This feature reflects the learning process when the user first enters the field,understands the background and overall knowledge of the field,and then learns the knowledge points in the field.(2)In the social space,the text abstraction does not show obvious changes when the text difficulty increases or decreases.This feature reflects that the difficulty of text in social space is not reflected in the form of abstraction.Combining the above conclusions,this paper formulates corresponding text content recommendation strategies for learning recommendations in the three spaces.Experimental results indicate that the recommendation strategy can effectively optimize the user's learning efficiency in the three spaces.