【摘要】：利用社會(huì)網(wǎng)絡(luò)分析技術(shù),構(gòu)建出游客社會(huì)網(wǎng)絡(luò),挖掘出游客間存在的局部社區(qū)關(guān)系,然后與傳統(tǒng)的協(xié)同過濾推薦算法進(jìn)行結(jié)合,能夠有效解決旅游產(chǎn)品推薦中數(shù)據(jù)稀疏性問題,另一方面向有社會(huì)關(guān)系的一群用戶推薦他們喜歡的旅游產(chǎn)品,能夠減少推薦的盲目性,提高推薦精準(zhǔn)度,并且能夠協(xié)助旅游公司給用戶提供更加個(gè)性化的服務(wù),提升旅游體驗(yàn)。為解決游客社會(huì)網(wǎng)絡(luò)構(gòu)建與社區(qū)發(fā)現(xiàn)問題,以真實(shí)的游客旅游記錄為基礎(chǔ),設(shè)計(jì)出了一種游客社會(huì)網(wǎng)絡(luò)構(gòu)建方法,并研究了基于中心節(jié)點(diǎn)擴(kuò)張的局部社區(qū)挖掘算法。該算法對PageRank算法進(jìn)行了修改,使其適用于社會(huì)網(wǎng)絡(luò)中節(jié)點(diǎn)的排名,在此基礎(chǔ)上研究了基于中心節(jié)點(diǎn)擴(kuò)張的局部社區(qū)挖掘算法。在基于社會(huì)網(wǎng)絡(luò)分析的旅游產(chǎn)品推薦算法中,通過計(jì)算同一局部社區(qū)內(nèi)用戶之間的直接信任度和間接信任度,量化為信任度值,然后對傳統(tǒng)的協(xié)同過濾推薦算法進(jìn)行改進(jìn),將用戶的信任度值融入到用戶相似性的計(jì)算當(dāng)中去。通過對某旅游公司的真實(shí)游客旅游記錄進(jìn)行加載、轉(zhuǎn)換和去噪,進(jìn)行實(shí)驗(yàn)。實(shí)驗(yàn)結(jié)果表明,本文提出的基于中心節(jié)點(diǎn)擴(kuò)張的局部社區(qū)挖掘算法可以有效地挖掘出游客社會(huì)網(wǎng)絡(luò)中存在的局部社區(qū),并且具有較小的時(shí)間復(fù)雜度。在推薦算法的對比實(shí)驗(yàn)當(dāng)中,本文采用平均絕對誤差MAE與準(zhǔn)確率作為對比,對比結(jié)果表明,本文提出的基于社會(huì)網(wǎng)絡(luò)分析的旅游產(chǎn)品推薦算法的MAE比傳統(tǒng)的協(xié)同過濾推薦算法降低了0.021,準(zhǔn)確率提高了2.5%。
[Abstract]:By using the social network analysis technology, the tourist social network can be constructed, the local community relationship among tourists can be excavated, and then combined with the traditional collaborative filtering recommendation algorithm, it can effectively solve the problem of data sparsity in tourism product recommendation. On the other hand, the recommendation of tourism products to a group of users with social relationship can reduce the blindness of recommendation, improve the accuracy of recommendation, and help travel companies to provide more personalized services to users and enhance the tourism experience. In order to solve the problem of social network construction and community discovery, a method of constructing tourist social network is designed based on the real tourist travel records, and the local community mining algorithm based on the expansion of central node is studied. The algorithm modifies the PageRank algorithm and makes it applicable to the ranking of nodes in the social network. On this basis, the local community mining algorithm based on the expansion of central nodes is studied. In the tourism product recommendation algorithm based on social network analysis, the direct trust and indirect trust between users in the same local community are calculated, and then the traditional collaborative filtering recommendation algorithm is improved. The trust value of the user is incorporated into the calculation of the user similarity. By loading, converting and de-noising the real tourist travel records of a tourism company, the experiment is carried out. The experimental results show that the proposed local community mining algorithm based on the expansion of central nodes can effectively mine the local communities existing in the tourist social network and has a relatively small time complexity. In the comparison experiment of the recommendation algorithm, the average absolute error (MAE) is compared with the accuracy rate. The comparison results show that, The MAE of tourism product recommendation algorithm based on social network analysis in this paper is 0.021 less than the traditional collaborative filtering recommendation algorithm, and the accuracy is improved 2.5%.
【學(xué)位授予單位】：中國民航大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2016
【分類號】：TP391.3
，

本文編號：2290514