本文關(guān)鍵詞： 茶產(chǎn)品 群組推薦 Phrase-LDA 融合策略　出處：《安徽農(nóng)業(yè)大學(xué)》2016年碩士論文　論文類型：學(xué)位論文

【摘要】：隨著網(wǎng)絡(luò)和信息技術(shù)的進(jìn)步,農(nóng)產(chǎn)品電子商務(wù)也飛速發(fā)展。根據(jù)國(guó)家商務(wù)部的數(shù)據(jù)顯示,2014年全國(guó)涉農(nóng)電子商務(wù)平臺(tái)正在蓬勃發(fā)展,目前已超三萬(wàn)家。以茶葉為代表的特色農(nóng)產(chǎn)品,也正在抓住機(jī)遇,從線下邁向線上,通過網(wǎng)絡(luò)走向千家萬(wàn)戶。由于涉農(nóng)電子商務(wù)的爆發(fā)式增長(zhǎng),網(wǎng)購(gòu)人群的迅速增加,如何在海量信息中找到符合用戶興趣的商品和如何向一群有相同興趣愛好的用戶推薦商品,成為亟待解決的問題。推薦系統(tǒng),利用消費(fèi)者在電商平臺(tái)上的隱性或者顯性的行為,分析其偏好,在過載的信息中,推薦滿足其偏好的商品。目前的推薦系統(tǒng)主要都是針對(duì)個(gè)人的,考慮用戶和商品數(shù)量的急劇增加,一對(duì)一的推薦的成本過高,如何向群組進(jìn)行推薦是非常的熱門研究方向。本文通過介紹現(xiàn)有群組推薦的相關(guān)理論,分析了目前電子商務(wù)環(huán)境下茶葉推薦存在的問題和基于評(píng)分推薦的不足,提出了基于Phrase-LDA模型的茶產(chǎn)品的群組推薦模型。具體研究?jī)?nèi)容如下:(1)基于茶產(chǎn)品的評(píng)論信息,提取用戶和茶產(chǎn)品評(píng)論中的主題,去表示用戶偏好和茶產(chǎn)品的主題特征。分析了目前傳統(tǒng)的用戶偏好的表示方法的優(yōu)點(diǎn)和不足,綜合其他方法的優(yōu)點(diǎn),提出主題信息表示法。(2)針對(duì)傳統(tǒng)的LDA是基于單詞的,單詞相對(duì)短語(yǔ)沒有明確的語(yǔ)義。本文提出Phrase-LDA,并用其去融合群組用戶偏好。Phrase-LDA主題模型一方面融合了群組用戶的偏好,另一方面在主題表示時(shí),短語(yǔ)的明確語(yǔ)義更細(xì)致的描述了用戶偏好。(3)利用京東商城部分茶葉的數(shù)據(jù)進(jìn)行實(shí)驗(yàn)仿真,驗(yàn)證所提Phrase-LDA主題模型的有效性,并實(shí)現(xiàn)了一個(gè)基于Phrase-LDA茶產(chǎn)品的群組推薦原型系統(tǒng)。本文的研究成果,進(jìn)一步細(xì)化了用戶偏好的研究。另外在群偏好的聚合策略方面,本文的方法有一定的新穎性,具有一定的借鑒意義。
[Abstract]:With the development of network and information technology, the rapid development of e-commerce of agricultural products. According to the national Ministry of commerce data show that in 2014 the national agricultural e-commerce platform is booming, currently has exceeded thirty thousand. Tea is the representative of the characteristics of agricultural products, is to seize the opportunity, from the line to line, through the network to thousands of households due to the explosive growth of agricultural e-commerce, online shopping population increased rapidly, how to find the mass of information goods and how to match the user's interest to a group of users have the same hobby to recommend commodities, has become an urgent problem. Recommendation system, the use of consumers in the electronic business platform of the implicit or explicit behavior and its analysis in preference, information overload, recommended to satisfy their preferences of goods. The current recommendation system is mainly for individual, considering the number of users and commodities sharply Increase, one of the recommended cost is too high, how to group recommendation is a popular research direction is. This paper introduces the related theory of existing group recommendation, analysis of the current e-commerce environment tearecommended problems and score recommended based on recommendation Phrase-LDA model based on the group of tea products specific contents are as follows: (1) the tea product reviews based on information extraction of user and tea product reviews the theme, theme features to represent user preferences and tea products. Analyzes the traditional user preference representation methods and the advantages and disadvantages, the comprehensive advantages of other methods, put forward the topic of information representation. (2) according to the traditional LDA is based on the word, the word phrase relative no clear semantics. In this paper Phrase-LDA, and used to merge user preferences in.Phrase-LDA model group On the one hand, the fusion group of user preferences, on the other hand, the theme said, clear semantic phrases more detailed description of the user preference. (3) simulation using part of the tea mall Jingdong data, verify the validity of the proposed Phrase-LDA model, and implements a prototype system of Phrase-LDA tea products recommended the group based on the results of this study, further refinement of the user preferences. In addition the group preference aggregation strategy, this method has a certain novelty, has a certain significance.

【學(xué)位授予單位】：安徽農(nóng)業(yè)大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2016
【分類號(hào)】：TP391.3

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 連彩琴;陳小婷;何靜;連雁平;;基于SEO茶葉電子商務(wù)推薦平臺(tái)優(yōu)化[J];福建電腦;2016年03期

2 張亮;;基于LDA主題模型的標(biāo)簽推薦方法研究[J];現(xiàn)代情報(bào);2016年02期

3 祝婷;秦春秀;馬曉悅;李祖海;;基于本體與LDA主題模型的文本資源推薦方法研究[J];情報(bào)雜志;2015年11期

4 劉佳佳;;消費(fèi)者網(wǎng)購(gòu)茶葉意愿影響因素分析[J];科技廣場(chǎng);2015年07期

5 王春梅;葉玉龍;劉躍云;陳敘生;;四川省宜賓市茶葉電子商務(wù)發(fā)展現(xiàn)狀及建議[J];安徽農(nóng)業(yè)科學(xué);2015年21期

6 焦東俊;;基于用戶人口統(tǒng)計(jì)與專家信任的協(xié)同過濾算法[J];計(jì)算機(jī)工程與科學(xué);2015年01期

7 鄭妍;龐琳;畢慧;劉瑋;程工;;基于情感主題模型的特征選擇方法[J];山東大學(xué)學(xué)報(bào)(理學(xué)版);2014年11期

8 歐衛(wèi);謝贊福;謝彬彬;歐繽憶;;基于LDA模型的社交網(wǎng)絡(luò)主題社區(qū)挖掘[J];計(jì)算機(jī)與現(xiàn)代化;2014年08期

9 李鵬;于曉洋;孫渤禹;;基于用戶群組行為分析的視頻推薦方法研究[J];電子與信息學(xué)報(bào);2014年06期

10 邸亮;杜永萍;;LDA模型在微博用戶推薦中的應(yīng)用[J];計(jì)算機(jī)工程;2014年05期

，

本文編號(hào)：1464745