Research And Implementation Of Entity Information Retrieval System Based On EPC Network

EPC (Electronic Product Code) network is a typical application ofInternet of things. EPC network is an open-loop framework using EPC asunique identifier to organize entity, storing and sharing information throughorganizational boundaries. Entity retrieval technology provides informationretrieval service in the perspective of the entity. Most of current researchesfocus on single product traceability and physical information security. Entityretrieval technology focus on terminal intelligent sensor deployment andcoordination. Represented by EPC network, frameworks for internet ofthings are all relatively weak to support information retrieval. And entityretrieval technology also needs to face to universal application scenario.In this thesis, EPC network architecture and entity information retrievaltechnology are in depth analyzed. Then we come up with the entityinformation retrieval solution based on EPC network. Firstly, the programextends EPC network architecture, adds internet information acquisitionmodule and entity information retrieval service layer which has three mainmodels that are entity statistics query engine, business analysis engine andentities search engine. Secondly, we put forward a content topic modelincludes five dimensions which are basic information, path information,status information, evaluation information and EPC information. The modelconstructs content topic information by improved similarity analysis algorithm, improves content retrieval efficiency by combining the localsensitivity hash method, and supports prediction of the states and analysisof evaluation information. Thirdly, the thesis proposes path encoding modelwhich can improve the efficiency of statistical path query, support efficientdata statistics across the nodes and support frequent path mining andprediction of path in future. Fourthly, data segmentation strategy is made toachieve distributed storage which has three major mechanism that are highsparse hash algorithm, two-layer structure of virtual server and physicalserver and adaptive management mechanism of free space and redundancy.Firstly, we review the EPC network architecture and entity retrievaltechnology. We deeply learn the characteristics of EPC network standardand the status of entity retrieval technology and summarize that it’s hard todo information retrieval across organizations. Secondly, after summarizingand analyzing the demand and challenges of entity information retrieval, wepropose proposed entity information retrieval system based EPC network-Gaia system. Thirdly, we come up with two major models and one methodswhich are content analysis model, path encoding model, data shardingstrategy. At last, we implement entity information retrieval system throughShandong Jingzhi wine product traceability project.