本文選題：網(wǎng)絡(luò)信息 + 個(gè)性化推薦系統(tǒng)�。� 參考：《長(zhǎng)安大學(xué)》2017年碩士論文

【摘要】：為了讓用戶從海量信息中高效地獲取自己需要的信息,推薦系統(tǒng)可以通過(guò)分析用戶歷史行為來(lái)了解用戶的偏好,從而主動(dòng)地為用戶推薦其感興趣的信息,滿足用戶的個(gè)性化需求。目前最近鄰思想、相似度思想和加權(quán)思想依然是各類推薦算法中最常用的思想,但隨著用戶數(shù)量和系統(tǒng)規(guī)模的不斷擴(kuò)大,推薦系統(tǒng)面臨著數(shù)據(jù)稀疏、流行偏置和可擴(kuò)展性差等問(wèn)題。傳統(tǒng)的推薦算法受數(shù)據(jù)稀疏的影響,相似度的準(zhǔn)確性不足,從而導(dǎo)致最近鄰搜索不準(zhǔn)確。本論文針對(duì)數(shù)據(jù)稀疏問(wèn)題,提出了基于奇異值分解的協(xié)同過(guò)濾改進(jìn)算法,該算法中用戶之間的相似度不再使用通用的評(píng)分矩陣計(jì)算,而是采用用戶特征向量矩陣計(jì)算。其具體實(shí)現(xiàn)方法是通過(guò)奇異值分解獲得用戶特征向量矩陣和物品特征向量矩陣,并利用用戶和物品之間的潛在關(guān)系,用奇異值去提取一些本質(zhì)特征,計(jì)算兩個(gè)用戶對(duì)應(yīng)特征向量的相似程度,從而得到用戶之間的相似度。此外,傳統(tǒng)的推薦算法往往傾向于推薦流行度較高的物品。本論文針對(duì)流行偏置問(wèn)題,在傳統(tǒng)的算法中根據(jù)物品流行度和用戶興趣信息引入懲罰函數(shù),從而提出了基于懲罰函數(shù)的協(xié)同過(guò)濾改進(jìn)算法。其算法的思想是從用戶行為數(shù)據(jù)的產(chǎn)生過(guò)程對(duì)用戶模型的影響進(jìn)行分析,根據(jù)物品流行度和用戶的興趣信息構(gòu)建懲罰度函數(shù),并使用懲罰函數(shù)調(diào)節(jié)不同流行度物品在用戶模型中權(quán)重。在MovieLens100K數(shù)據(jù)集上,對(duì)面向稀疏矩陣偏置的協(xié)同過(guò)濾推薦算法性能進(jìn)行了驗(yàn)證,實(shí)驗(yàn)結(jié)果表明基于奇異值分解的協(xié)同過(guò)濾改進(jìn)算法比基于用戶的協(xié)同過(guò)濾推薦算法的準(zhǔn)確率提升了0.68%,改善了最近鄰搜索的準(zhǔn)確性,其適用于新聞推薦領(lǐng)域�；趹土P度的協(xié)同過(guò)濾推薦算法比基于用戶的協(xié)同過(guò)濾推薦算法的覆蓋率提高了1.47%,緩解了流行度對(duì)用戶行為的影響,其適合電子商務(wù)推薦領(lǐng)域。
[Abstract]:In order for users to obtain the information they need efficiently from the mass of information, the recommendation system can analyze the historical behavior of users to understand the preferences of users, so as to actively recommend the information of interest to the users. To meet the personalized needs of users. At present, the nearest neighbor thought, the similarity thought and the weighted thought are still the most commonly used ideas in all kinds of recommendation algorithms, but with the continuous expansion of the number of users and the scale of the system, the recommendation system is faced with sparse data. Popular bias and poor scalability and other problems. The traditional recommendation algorithm is influenced by sparse data, and the accuracy of similarity is insufficient, which leads to inaccuracy of nearest neighbor search. In this paper, an improved collaborative filtering algorithm based on singular value decomposition (SVD) is proposed to solve the data sparsity problem. In this algorithm, the similarity between users is calculated not by the common score matrix, but by the user eigenvector matrix. The implementation method is to obtain user eigenvector matrix and item eigenvector matrix by singular value decomposition, and extract some essential features by singular value using the latent relation between user and item. The similarity between the two users is obtained by calculating the similarity of the corresponding feature vectors between the two users. In addition, traditional recommendation algorithms tend to recommend items with high popularity. Aiming at the problem of popular bias, this paper introduces a penalty function according to the information of item popularity and user's interest in the traditional algorithm, and then proposes an improved collaborative filtering algorithm based on penalty function. The idea of the algorithm is to analyze the influence of the user behavior data on the user model, and to construct the penalty degree function according to the item popularity and the user's interest information. And the penalty function is used to adjust the weight of different popular items in the user model. Based on MovieLens100K data set, the performance of collaborative filtering recommendation algorithm for sparse matrix bias is verified. The experimental results show that the improved collaborative filtering algorithm based on singular value decomposition improves the accuracy of collaborative filtering recommendation algorithm by 0.68, improves the accuracy of nearest neighbor search, and is suitable for news recommendation field. Compared with the user-based collaborative filtering recommendation algorithm, the penalty degree based collaborative filtering recommendation algorithm increases the coverage by 1.47%, which alleviates the influence of popularity on user behavior and is suitable for e-commerce recommendation field.
【學(xué)位授予單位】：長(zhǎng)安大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TP391.3

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 韓亞楠;曹菡;劉亮亮;;基于評(píng)分矩陣填充與用戶興趣的協(xié)同過(guò)濾推薦算法[J];計(jì)算機(jī)工程;2016年01期

2 榮輝桂;火生旭;胡春華;莫進(jìn)俠;;基于用戶相似度的協(xié)同過(guò)濾推薦算法[J];通信學(xué)報(bào);2014年02期

3 孫光福;吳樂(lè);劉淇;朱琛;陳恩紅;;基于時(shí)序行為的協(xié)同過(guò)濾推薦算法[J];軟件學(xué)報(bào);2013年11期

4 徐風(fēng)苓;孟祥武;王立才;;基于移動(dòng)用戶上下文相似度的協(xié)同過(guò)濾推薦算法[J];電子與信息學(xué)報(bào);2011年11期

5 李改;李磊;;基于矩陣分解的協(xié)同過(guò)濾算法[J];計(jì)算機(jī)工程與應(yīng)用;2011年30期

6 馬宏偉;張光衛(wèi);李鵬;;協(xié)同過(guò)濾推薦算法綜述[J];小型微型計(jì)算機(jī)系統(tǒng);2009年07期

7 肖敏;熊前興;;基于項(xiàng)目語(yǔ)義相似度的協(xié)同過(guò)濾推薦算法[J];武漢理工大學(xué)學(xué)報(bào);2009年03期

8 彭德巍;胡斌;;一種基于用戶特征和時(shí)間的協(xié)同過(guò)濾算法[J];武漢理工大學(xué)學(xué)報(bào);2009年03期

9 彭玉;程小平;;基于屬性相似性的Item-based協(xié)同過(guò)濾算法[J];計(jì)算機(jī)工程與應(yīng)用;2007年14期

10 余力,劉魯;電子商務(wù)個(gè)性化推薦研究[J];計(jì)算機(jī)集成制造系統(tǒng);2004年10期

相關(guān)博士學(xué)位論文 前2條

1 孔維梁;協(xié)同過(guò)濾推薦系統(tǒng)關(guān)鍵問(wèn)題研究[D];華中師范大學(xué);2013年

2 孫小華;協(xié)同過(guò)濾系統(tǒng)的稀疏性與冷啟動(dòng)問(wèn)題研究[D];浙江大學(xué);2005年

相關(guān)碩士學(xué)位論文 前6條

1 呼亞杰;一種基于類別偏好協(xié)同過(guò)濾推薦算法的實(shí)現(xiàn)與優(yōu)化[D];蘭州大學(xué);2016年

2 張磊;基于改進(jìn)的K-means和奇異值分解的協(xié)同過(guò)濾研究[D];大連海事大學(xué);2015年

3 陳清浩;基于SVD的協(xié)同過(guò)濾推薦算法研究[D];西南交通大學(xué);2015年

4 �？箍�;流行度對(duì)用戶興趣的影響機(jī)制分析及其在推薦算法中的應(yīng)用研究[D];上海大學(xué);2015年

5 周超;面向數(shù)據(jù)稀疏問(wèn)題的協(xié)同過(guò)濾推薦算法改進(jìn)研究[D];杭州電子科技大學(xué);2013年

6 陳登科;基于潛在向量模型與項(xiàng)目的協(xié)同過(guò)濾混合推薦[D];浙江大學(xué);2010年

，

本文編號(hào)：1776099