本文選題：個性化推薦 + 用戶標(biāo)注��； 參考：《電子科技大學(xué)》2012年碩士論文

【摘要】：隨著現(xiàn)代電子商務(wù)和網(wǎng)絡(luò)的快速發(fā)展，各大商務(wù)網(wǎng)站都為用戶提供了越來越人性化個性化的服務(wù)，個性化的推薦搜索引擎的研究也越來越廣泛的被應(yīng)用。目前各大商務(wù)網(wǎng)站的個性化推薦引擎基本都是統(tǒng)一推薦或者是針對熱門類目進(jìn)行推薦，沒有設(shè)計出一種針對不同的用戶給出不同的推薦信息的推薦搜索引擎，這種搜索引擎就像是了解你性格興趣的商品推銷員，能夠總是在你需要某些商品信息的時候第一時間把推薦展示在你面前。 本文針對上述問題，在分析和研究主流的搜索引擎的基礎(chǔ)上設(shè)計出一個簡易的支持短文本的個性化推薦搜索引擎系統(tǒng)包括OnLine模塊和OffLine模塊。本系統(tǒng)能夠根據(jù)不同的注冊用戶，并針對該用戶的興趣愛好給出個性化的推薦信息。本系統(tǒng)研究的主要內(nèi)容為： 1.在線OnLine處理模塊，其中包括Http服務(wù)器，query分析器和用戶特征獲取器，Rank評分核心機制，排序評分排序等幾個部分。設(shè)計一個簡單Http服務(wù)器來作為本系統(tǒng)的一個服務(wù)器容器，由于本文研究設(shè)計的個性化推薦搜索引擎系統(tǒng)是一個輕量級的系統(tǒng)，因此需要一個同樣簡易化輕量級的Http網(wǎng)絡(luò)服務(wù)器來支持。通過用戶特征獲取器來獲取該用戶的基本信息和興趣愛好，query分析器用來獲取用戶查詢記錄中相關(guān)記錄處理后的一個倒排表。Rank評分核心機制，也是本系統(tǒng)的核心，對查詢分析和用戶特征獲取器獲取的所有數(shù)據(jù)記錄進(jìn)行評分處理，依據(jù)分?jǐn)?shù)排序，獲取一定數(shù)量排在前面的結(jié)果集，即是根據(jù)用戶的興趣愛好和購買歷史所推薦的優(yōu)先結(jié)果集，優(yōu)先返回和用戶興趣愛好相關(guān)性較大的記錄。 2.離線OffLine模塊，該模塊主要負(fù)責(zé)后臺數(shù)據(jù)的處理。后臺需要處理的數(shù)據(jù)主要是用戶標(biāo)注模塊和查詢標(biāo)注模塊。用戶信息分為基本信息和行為信息，用戶的基本信息通過實驗數(shù)據(jù)的方式獲得，基本信息中包含用戶的最基本的特征，，而用戶的行為信息中一般分為興趣和購買歷史兩個部分，這些信息反映了用戶的興趣愛好等特征，通過對用戶的信息分析和用戶查詢記錄的分析，過濾對本次查詢無效的記錄。
[Abstract]:With the rapid development of modern electronic commerce and network, each major business website provides more and more personalized service for users, and the research of personalized recommendation search engine is more and more widely used. At present, the personalized recommendation engine of each major business website is basically a unified recommendation or a recommendation for hot categories. There is no design of a recommendation search engine for different users to give different recommendation information. This search engine is like a merchandiser who understands your personality interests and can always present recommendations to you the first time you need information about something. In this paper, based on the analysis and research of the mainstream search engine, a simple personalized recommendation search engine system supporting short text is designed, which includes OnLine module and OffLine module. The system can give personalized recommendation information according to different registered users and their interests. The main contents of this system are as follows: 1. Online OnLine processing module, which includes Http server query analyzer and user feature acquirer Rank scoring core mechanism, ranking rating ranking and other parts. A simple Http server is designed as a server container of this system. Because the personalized recommendation search engine system studied in this paper is a lightweight system, Hence the need for an equally easy lightweight Http web server to support. The basic information and interests of the user are obtained by the user feature acquirer. The query analyzer is used to obtain an inverted table. Rank scoring core mechanism after the processing of the related records in the user query record, which is also the core of the system. All the data records obtained by the query analysis and the user feature acquirer are graded, and according to the ranking of the scores, a certain number of the first result sets are obtained, that is, the priority result set recommended according to the user's interest and purchase history. Priority return and user interests related to the record. 2. Offline OffLine module, this module is mainly responsible for background data processing. The data needed to be processed in the background are the user tagging module and the query annotation module. User information is divided into basic information and behavior information, the basic information of users is obtained through experimental data, the basic information contains the most basic characteristics of users, and the behavior information of users is generally divided into two parts: interest and purchase history. This information reflects the characteristics of the user's interests and hobbies. Through the analysis of the user's information and the analysis of the user's query records, the invalid records of this query are filtered.
【學(xué)位授予單位】：電子科技大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2012
【分類號】：TP391.3

【參考文獻(xiàn)】

相關(guān)期刊論文 前5條

1 雷鳴,王建勇,趙江華,單松巍,陳葆玨;第三代搜索引擎與天網(wǎng)二期[J];北京大學(xué)學(xué)報(自然科學(xué)版);2001年05期

2 紀(jì)良浩;王國胤;楊勇;;基于協(xié)作過濾的Web日志數(shù)據(jù)預(yù)處理研究[J];重慶郵電學(xué)院學(xué)報(自然科學(xué)版);2006年05期

3 賀衛(wèi)紅;曹毅;;基于向量空間模型文本過濾算法[J];系統(tǒng)工程;2005年10期

4 宋聚平,王永成;搜索引擎中信息存儲結(jié)構(gòu)的改進(jìn)[J];情報學(xué)報;2001年05期

5 黃羿;馬新強;鄒永康;;用戶興趣模型在個性化搜索引擎中的應(yīng)用[J];重慶文理學(xué)院學(xué)報(自然科學(xué)版);2008年05期

相關(guān)碩士學(xué)位論文 前2條

1 羅勝陽;協(xié)同過濾技術(shù)及其在電子商務(wù)推薦領(lǐng)域的應(yīng)用研究[D];南京理工大學(xué);2008年

2 譚雷雨;科技文獻(xiàn)共享平臺全文檢索技術(shù)研究[D];華中科技大學(xué);2007年

本文編號：1783731