Research On Sleep Knowledge Automatic Classification Algorithm Based On Subject Map

With the explosive growth of information resources,traditional knowledge getting method cannot meet the demand of users in the hurge data environment.We should make revolution of learning method,to create new mode of medical information service and solve big data.At the request of the new era,people want an understanding domain knowledge,shape the domain knowledge contour,and enter the mobile software that can help people master the domain knowledge.In order to achieve the goal,our first task is to obtain the target knowledge accurately from the vast amount of Internet information resources.Text classification is the key technology to solve this problem.However,the traditional text classification model has great limitations,cannot analyze the semantic relation between the keywords in the document,nor can it meet the basic needs of hierarchical classification.However,the traditional text classification model has great limitations,cannot analyze the semantic relation between the keywords in the document,nor can it meet the basic needs of hierarchical classification.In order to solve this problem,this paper proposes a text categorization algorithm based on a single subject knowledge map,and on the basis of this algorithm builds a sleep subject knowledge service system.In order to achieve the purpose of semantic network hierarchical classification,this paper introduces the knowledge mapping technology.First,we create the knowledge map in a single subject area,then we combine the map and the classical SVM(Support Vector Machine)text classification model to analyse the similarity of text and knowledge entities.This is the base of our subject knowledge service system.The research contents of this paper mainly divide into three aspects:Use encyclopedia data to construct the knowledge map.First,this paper researched and analyse the seed of encyclopedia data,and structured the seed corpus;then this paper used the Chinese word segmentation,entity recognition technologies to extract entities;then this article use machine learning algorithms,two way LSTM(Long Short-Term Memory)to extract the relation betweent entities.At last,we have created the knowledge map and tested the result of experiment.This paper propose a text classification algorithm which is based on the subject map.The main idea of the algorithm is to divide the text classification task into smaller sub-problem by the construction of the theme ma,improveing the efficiency of text classification.Given that the topic map of this paper is a non-loop-type tree structure,we have adopted a top-down text classification model.Specifically,according to each node in the topic map sample subset GeXun build each classifier,and then the top-down by the classifier to classify the large text entities assigned to the topic map categories,so as to realize the semantic network of text classification.Finally,based on the algorithm and the theme map,this paper build a sleep subject knowledge service system,to support the mobile devices,subject knowledge automatic acquisition and push responsible for sleep.The main structure of the system is: firstly,the topic map is used to effectively mine and analyze the knowledge data in the field of expertise,and use the text classification algorithm to search the relevant knowledge of the topic;Thereafter,the knowledge system is constructed with each entity as the information organization unit and the unit information as the hub.Finally,users are provided with mobile devices.