Full metadata record
DC FieldValueLanguage
dc.contributor.author廖建智en_US
dc.contributor.author孫春在en_US
dc.date.accessioned2014-12-12T02:25:11Z-
dc.date.available2014-12-12T02:25:11Z-
dc.date.issued2000en_US
dc.identifier.urihttp://140.113.39.130/cdrfb3/record/nctu/#NT890394064en_US
dc.identifier.urihttp://hdl.handle.net/11536/66967-
dc.description.abstract網路盛行以來,資訊的取得更為容易,Napster流行以後,網路上的檔案分享更是達到前有未有的盛況,但是卻也因此造成「資訊超載」的情況,讓人們難以從龐大的資料海中找到自己需要的資訊。在Napster之後,這種分散式資源分享工具如雨後春筍般的出現,但是絕大部分的分享工具只在強調自己的分享族群有多麼龐大,而對於如何讓使用者應付這龐大分享族群所帶來的資料量卻著墨甚少。因此,在本篇論文中以分散式資源分享系統為對象,以使用者興趣相近與否為依據,建立社會網路 (social network),並利用這個網路的特性,為搜尋結合合作式資訊過濾以增加搜尋的精確度,同時加入推薦的功能,利用同好的資訊過濾成果以節省自己過濾資訊的精力,增加面對龐大資料量的處理能力。 在本篇論文中,實作了一個非中控分散式資源分享系統,並在其中實現搜尋、推薦、網路自我調整的能力。並且設計一連串的模擬實驗,驗證這個系統各種功能的表現,看看是否加入同好網路和推薦功能後,能讓使用者獲得更多他們所需的資源。最後探討在實驗中觀察到的現象和問題,並提出可能的改進方法,讓系統和使用者都能更有效率的分享交換資源。zh_TW
dc.description.abstractWith the growth of Internet, accessing information is more convinent than past. File sharing becomes a popular application on Internet after the development of Napster. On the other side, the increase of accessing information raises the ‘information overloading’ problem, which means that people are hard to locate the useful information within enormous resources. The distributed file sharing applications emerge right after Napster is proposed. Most of them do not provide the solution helping people to handle enormous information but place emphasis on the scale of the sharing community. In this paper, we discuss the ‘information overloading’ issue on distributed file sharing systems. We construct a ‘social network’ based on the preference similarity among users, and utilize the features of this network to combine search function with collaborative information filtering method. Besides, we have applid an additional mechanism of recommendation on the system in order to help people handling huge amount of information. In this paper, we proposed a decentralized distributed resource sharing system, and implemented all the functions of search, recommendation and network self-organizing in it. In order to verify the performance of this system, we designed a series of simulations with distinct conditions. We also observe if users can obtain more resources they want after they use this system. Finally, we discuss some problems arisen from the simulations, and propose solutions to improve the efficiency of resources sharing and exchanging.en_US
dc.language.isozh_TWen_US
dc.subject分散式zh_TW
dc.subject資源分享zh_TW
dc.subject搜尋zh_TW
dc.subject推薦zh_TW
dc.subject檔案分享zh_TW
dc.subject資訊擷取zh_TW
dc.subject資訊過濾zh_TW
dc.subject自我調整zh_TW
dc.title分散式資源分享與推薦zh_TW
dc.typeThesisen_US
dc.contributor.department資訊科學與工程研究所zh_TW
Appears in Collections:Thesis