The Research And Implementation Of Assistant Reading System

In recent years,with the rapid development of the Internet and big data,users often need to open multiple pages to find relevant knowledge to help them understand when reading electronic articles.Most current reading auxiliary system only interpretation on the part of the article,it exists the shortage of the following three aspects:(1)it to the user how to quickly get less can help understand the content of the help;(2)for different users interested in the same article content is different,it is difficult to provide targeted services;(3)it ignores the memory for the user to read and understand the reading order.This article through to the user interest model is established to describe the user's individualized demand,through the interest model and read the article,recommended to the user can help understand the text content and reading sequence,and by reference to the literature and the related articles to read content expanded,expanding the scope of the proposed text,further to help users to understand the text.Based on the above work,this article implements an auxiliary reading system.The specific contents of this article are as follows:(1)Build interest model.This article builds an interest model for users based on the relevant information provided by users.In WordNet corpus building methods use noun fluctuation relationship and interest in the key words in the context of information,using the keyword weight(tf-idf,Term Frequency,Inverse Document Frequency)model for interest interest points in priority order with indicators,take the shape of node weighted directed graph model reflect users' interests gradation and links,and mining interests in the model more closely interest community as a measure of interest model.(2)Recommended reading orientation.We propose a mapping method by analyzing semantic similarity and structural similarity.This method turns the reading articles into a topic network,and then compares and analyzes the similarity between the network and the interest community in the interest model to determine the recommended reading area,and then recommends users to read according to the relevance between the interest community and the node of the topic network.(3)Expand the content.We use vector to represent the topic in the reading content,then use vector matching to retrieve the paragraphs related to the reading content in the references or references related to the reading article,and finally present the matching results to users after comprehensive,de-duplication and other operations.The innovation of this article mainly includes:(1)Improve the traditional reading system.We focus on user interest expression,reading positioning and reading expansion,quickly determine the content that users are interested in reading articles,and expand the reading content to help users more comprehensively understand the reading content.Compared with the traditional reading system,this article focuses on the acquisition and understanding of reading content.(2)Improvements to the reading process.We reconstruct the reading order of articles based on the interest points in the interest community,and recommend users to read the content related to interest first.This reading sequence helps users form serial memory,deepen their understanding of the reading content,and reduce the repetition of reading.(3)From text recommendation to content recommendation.The general text recommendation method only recommends the text,and readers still need to identify the content they are interested in or need to read in the text.The reading assistance system designed in this article can recommend the contents of articles according to users' interests,helping users to get the target contents of texts more quickly.