The Design And Implementation Of Semi-automatic Retrival Based On QGS In SLR System

Systematic Literature Review(SLR)method is one of the most common research methods in Evidence-Based Software Engineering.The researcher who uses SLR method needs to collect amounts of data on the Internet.Quasi-Gold Standard(QGS)is a quasi-gold standard to evaluate the rationality of the collected data which is able to assist researchers in completing the stage of literature search.However,researchers need to spend lots of time when they use SLR method to conduct research and the data generated in the research process is difficult to save.In this regard,our laboratory develops an SLR system,aiming to help researchers using this method.The system covers various phases of the SLR method and records process data for researchers.Literature search is the first stage to implement SLR,its retrieved result can influence and even determine the success or failure of the project,whereas the literature retrieval process based on QGS is very time-consuming.There is no online collaborative tools released helping researchers complete literature retrieval based on the concept of QGS at present.The literature search module discussed in this thesis provides researchers with literature retrieval function based on QGS.This thesis describes the background of the SLR system,sets out semi-automatic retrival based on QGS in which played an important role,summarizes the development of some assistive tools in academic research and introduces technique used in semi-automatic retrival based on QGS such as Django,Redis and the theme web crawler.After analysing the system requirements in detail,the thesis depicts the overall struc-tural design of the system,the module and the database.The module is subdivided into four sub-modules:agent pool module,digital library crawler module,information retrieval module and data export module.The detailed design and implementation of the four sub-modules are given from the functional level in this thesis.Finally,this thesis conducts tests targeted on several main modules.In order to solve the problem that the same search string has different semantics in different digital libraries,this thesis presents and implements search string format-ting,achieving semantic consistency.The system introduces into csv file retrieval and parsing to decrease the number of high HTTP requests inherent to crawlers compared to other vertical search engines based on theme web crawler,reducing time consump-tion.In addition,year interval retrieval,an optimization strategy further reduces the time cost of retrieval.In the end,the unique function of calculating QS value provides strong support for researchers using SLR method.