Intelligent Data Organization Method And Traffic Application

In the 21st century,human society has fully entered the digital era.With the explosion of data,the storage and application methods of data bring many new challenges to problem solving.Among them,it is inevitable to generate arguments through "phishing" argument in big data,and the core reason is the chaotic organization method of big data.Therefore,this paper studies and proposes a "data intelligence" framework to realize the intelligent organization of data and task solving,and carries out a prototype application of text task processing in the transportation industry.The main work of this paper includes:Based on the analysis of the current situation,this paper proposes a "data intelligence" framework based on inclusion architecture,which is divided into data production layer,data organization layer,data collaboration layer and data application layer.In the aspect of data organization,a new data set description method is designed through feature analysis,and the encapsulation,storage and implementation method of general data set is given based on object storage technology.Based on the analysis of data collaboration and application,the basic process of text task processing using data intelligence is proposed,including text task parsing,data agent matching and answer extraction.Secondly,a multi-method text task parsing model was designed and implemented.Text tasks are presented in the form of natural language problems and are divided into four categories according to their solutions.For object retrieval and content extraction tasks,the task parsing model makes category annotation and entity annotation on the problem text.The Bi LSTM-CRF model based on attention mechanism was used for question entity recognition in the process of text task parsing.An improved Attention-Based Text CRNN model was proposed to recognize the type of text tasks.Furthermore,according to the parsing results of text tasks,a model of active matching between task questions and text data agents and autonomous answer extraction is designed and implemented.Based on the keyword extraction of the text data agent,the cosine similarity of the keywords between the task text and the document is calculated based on the Word2 Vec model,and the matching between the text data agent and the task question is established.This paper proposes an answer extraction model based on LDA model and improved Bi DAF.By calculating the topic probability and joint probability of the document,the answer sentence extraction of multiple paragraphs is realized.Finally,an intelligent document management prototype system for transportation industry applications was designed and implemented.The requirement analysis and architecture design of the system are carried out,and the function implementation and prototype of the main modules in the system are completed,which proves the feasibility of the task processing based on the data intelligence framework proposed in this paper.