【摘要】：隨著互聯(lián)網(wǎng)技術(shù)的蓬勃發(fā)展,網(wǎng)上信息資源的數(shù)據(jù)量也呈現(xiàn)出爆發(fā)式增長。在農(nóng)資交易平臺中,農(nóng)戶如何從大量農(nóng)資商品中找到符合自己需求的商品,以及供應(yīng)商如何讓自己的商品脫穎而出,都是一件非常困難的事情。目前,針對不同角色用戶,不同地域用戶及用戶在不同季節(jié)的需求特性,提供滿足其個性化要求的信息服務(wù),已成為農(nóng)資電子商務(wù)站點亟需解決的難題。協(xié)同過濾推薦系統(tǒng)作為一種重要的個性化服務(wù)模式,在互聯(lián)網(wǎng)領(lǐng)域的應(yīng)用越來越廣泛。本文以農(nóng)藥為例,研究出一種基于用戶角色的推薦算法,與傳統(tǒng)的推薦算法相比,此種算法綜合了農(nóng)資季節(jié)性,地域性,使用特性等特點,更適用于農(nóng)資推薦。同時,本文將智能推薦技術(shù)與農(nóng)資交易平臺相結(jié)合,設(shè)計實現(xiàn)基于用戶角色的農(nóng)資供求信息智能推薦系統(tǒng)。本文的主要研究內(nèi)容如下:現(xiàn)有推薦系統(tǒng)及推薦算法發(fā)展現(xiàn)狀研究。對基于內(nèi)容的推薦算法、協(xié)同過濾推薦算法等相關(guān)基礎(chǔ)理論進行了較為深入的研究,結(jié)合農(nóng)資電商推薦系統(tǒng)現(xiàn)狀,對如何構(gòu)建適用農(nóng)資交易平臺的個性化推薦算法進行了深入的研究分析。改進的協(xié)同過濾算法研究。通過計算I-I相似矩陣及收集用戶隱式行為,建立修正的I-U評分矩陣。并構(gòu)造用戶相似度矩陣U-U及Pearson相關(guān)系數(shù)計算用戶相似性,確定最近鄰用戶,最后生成預(yù)測評分及推薦項。農(nóng)資供求信息智能推薦系統(tǒng)的實現(xiàn)。在建立了實體間的E-R圖基礎(chǔ)上,對數(shù)據(jù)表進行詳細設(shè)計。搭建storm分布式實時計算框架,設(shè)計開發(fā)集農(nóng)資購買、個性推薦、訂單管理、和基礎(chǔ)查詢?yōu)橐惑w的農(nóng)資供求信息智能推薦系統(tǒng)。在安徽省“十二五”科技攻關(guān)項目課題“面向全程電子商務(wù)的農(nóng)資物流信息化關(guān)鍵技術(shù)研發(fā)與應(yīng)用”的支持下,本論文研究成果成功地將基于用戶角色的協(xié)同過濾推薦算法運用到農(nóng)資產(chǎn)品個性化推薦的服務(wù)中,有效地減少用戶的搜索時間,促進了交易的完成。
[Abstract]:With the rapid development of Internet technology, the amount of data of information resources on the Internet also presents explosive growth. In the agricultural material trading platform, it is very difficult for farmers to find the goods that meet their needs from a large number of agricultural products and how suppliers make their commodities stand out. At present, according to the demand characteristics of different users, different regions and users in different seasons, providing information services to meet their personalized requirements has become a difficult problem that needs to be solved in e-commerce sites of agricultural materials. Collaborative filtering recommendation system, as an important personalized service model, is more and more widely used in the field of Internet. In this paper, a recommendation algorithm based on user's role is developed by taking pesticide as an example. Compared with the traditional recommendation algorithm, this algorithm combines the characteristics of seasonality, regionality and use characteristics of agricultural materials, and is more suitable for recommendation of agricultural materials. At the same time, this paper combines the intelligent recommendation technology with the agricultural material trading platform to design and implement the intelligent recommendation system based on the user role of agricultural material supply and demand information. The main contents of this paper are as follows: the current status of recommendation systems and recommendation algorithms. Based on the content of the recommendation algorithm, collaborative filtering recommendation algorithm and other related basic theory is more in-depth research, combined with the status quo of agricultural e-commerce recommendation system, This paper analyzes how to construct personalized recommendation algorithm which is suitable for agricultural material trading platform. Research on improved collaborative filtering algorithm. By calculating I-I similarity matrix and collecting implicit behavior of users, a modified I-U score matrix is established. The user similarity matrix U-U and Pearson correlation coefficient are constructed to calculate the user similarity, and the nearest neighbor user is determined. Finally, the prediction score and recommendation items are generated. The realization of the intelligent recommendation system of agricultural supply and demand information. Based on the establishment of E-R graph between entities, the data table is designed in detail. A storm distributed real-time computing framework is built to design and develop an intelligent recommendation system for agricultural material supply and demand information, which integrates agricultural material purchase, individual recommendation, order management and basic query. Supported by the "12th Five-Year Plan" key Science and Technology Project in Anhui Province, "Research, development and application of key technologies for agricultural material logistics information oriented to the whole process of electronic commerce", In this paper, the collaborative filtering recommendation algorithm based on user role is successfully applied to the personalized recommendation service of agricultural products, which can effectively reduce the search time of users and promote the completion of transactions.
【學(xué)位授予單位】：安徽農(nóng)業(yè)大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2016
【分類號】：TP391.3

【參考文獻】

相關(guān)期刊論文 前10條

1 龔燦;盧軍;;基于Spark的實時情境推薦系統(tǒng)關(guān)鍵技術(shù)研究[J];電子測試;2016年Z1期

2 王春才;邢暉;李英韜;;個性化推薦系統(tǒng)冷啟動問題研究[J];現(xiàn)代計算機(專業(yè)版);2015年29期

3 彭明喜;;基于Storm的實時數(shù)據(jù)平臺研究[J];電信快報;2015年07期

4 張獻忠;;基于商品流行度和用戶活躍度的推薦算法研究[J];電腦知識與技術(shù);2014年32期

5 蔣勝;王忠群;修宇;皇蘇斌;;集成社會化標簽和用戶背景信息的協(xié)同過濾推薦方法[J];計算機應(yīng)用;2014年08期

6 文俊浩;舒珊;;一種改進相似性度量的協(xié)同過濾推薦算法[J];計算機科學(xué);2014年05期

7 陳龍飛;趙雪;;關(guān)于農(nóng)資產(chǎn)品個性化推薦的研究[J];電子世界;2013年24期

8 于洪;李俊華;;結(jié)合社交與標簽信息的協(xié)同過濾推薦算法[J];小型微型計算機系統(tǒng);2013年11期

9 劉文龍;張桂蕓;陳U，

本文編號：2202495