以任務佇列模式設計整合替代計量分析系統

Abstract

替代計量學（altmetrics）是文獻計量分析領域的新興研究領域，用以取代傳統書目計量學（bibliometrics）與資訊計量學（informetrics）在資料來源上的限制，其整合了網路計量學的優點與特性，將資料來源擴展至許多非正式的學術活動行為，讓學者的各項網路活動、討論交流、課程與程式產出等等也能列入統計。可以用來補足傳統計量學只能研究專業資料庫所收錄之文獻，且更新與散播速度較慢的問題。雖然國外已經有許多 Altmetrics 分析工具，但這些工具對國內在地化研究資訊分析能力的不足，需要一個能夠以中文作者姓名來分析其著作影響力的系統。本研究旨在探索 Altmetrics 適合用於國內學者網路影響力的資料來源與類型，並根據這些分類打造中文的替代計量分析系統，補充國外替代計量分析系統的不足，作為探討學者與其著作影響力的參考資源。由於此分析系統必須適應各類不同的網站平台，並抓取對應的資料作分析，而這些資料都是非結構化的，且分析的規則也各不相同。為此必須設計一套能夠適應多種環境，並且容易擴充新資料來源的架構。在本研究中，將採用任務佇列模式，藉由多組任務列隊依次執行並指派新任務，來達到彈性擴充的目的。最後所設計出來的系統擁有優良的GUI介面，並且分析內容與人工分析的差異不大，有一定程度的參考價值，足以取代大量人工查詢的工作。在未來，期許能有更多的研究，在本研究之上，設計出更多的資料分析類型，與更精確的分析演算法，讓國內學者的努力被更多人看見。

Altmetrics is a new research field which to resolve the limit of traditional Bibliometrics and Informetrics. It’s integrated benefits and features from Webometrics, extends the data source from many informal activities of academic and made all activities of internet, discussion, courses and code that can be included to the statistic. It can be used to solve the problem of traditional Informetrics which is too slow on spreading and only includes data from professional database. Although there are many altmetrics tools in foreign country, but they are limited on analysis local information, we need a new system to analysis the influences of scholars through Chinese name. The subject of this thesis is that to explore the data source and data type to analysis the influence of local scholars, design a Chinese Altmetrics analysis system based on the result of research and covers the shortage of the tools in foreign country on researching local scholars. This system must adapt many different websites/platforms and fetch corresponding data, but these data are unstructured, and must analysis by different rules. We must design a structure which is able to adapt to many environments and easily to extend new data source. In this thesis, will use task queue pattern, execute and assign tasks by many tasks in a list. In the result, we created a system with well GUI, and there are very few differences between system auto analysis and manual analysis. It has something referential value and is available to replace the manual analysis. We can expect there will be more research to find more data type and develop more algorithm to make the effort of local scholars be known by more people.