基于收視行為的互聯(lián)網(wǎng)電視節(jié)目流行度預(yù)測(cè)模型
doi: 10.11999/JEIT161310
基金項(xiàng)目:
國(guó)家計(jì)劃項(xiàng)目(863)(2015AA015603)，江蘇省未來(lái)網(wǎng)絡(luò)創(chuàng)新研究院未來(lái)網(wǎng)絡(luò)前瞻性研究項(xiàng)目(BY2013095-5-03)，江蘇省六大人才高峰高層次人才項(xiàng)目(2011-DZ024)
Program Popularity Prediction Model of Internet TV Based on Viewing Behavior
Funds:
The National 863 Program of China (2015AA 015603), The Prospective Research Program on Future Networks of Jiangsu Province (BY2013095-5-03), The Six Industries Talent Peaks Plan of Jiangsu Province (2011-DZ024)
-
摘要: 準(zhǔn)確預(yù)測(cè)節(jié)目流行度是互聯(lián)網(wǎng)電視節(jié)目系統(tǒng)設(shè)計(jì)與優(yōu)化所要解決的關(guān)鍵問(wèn)題之一。針對(duì)現(xiàn)有預(yù)測(cè)方法存在模型訓(xùn)練時(shí)間長(zhǎng)、樣本數(shù)量多、且對(duì)突發(fā)熱點(diǎn)節(jié)目流行度預(yù)測(cè)效果差等問(wèn)題，該文測(cè)量了某互聯(lián)網(wǎng)電視平臺(tái)280萬(wàn)用戶(hù)的60億條收視行為數(shù)據(jù)，采用行為動(dòng)力學(xué)分類(lèi)方法將節(jié)目流行度演化過(guò)程分為內(nèi)源臨界、內(nèi)源亞臨界、外源臨界和外源亞臨界4種類(lèi)型，運(yùn)用雙種群粒子優(yōu)化的最小二乘支持向量機(jī)對(duì)每種類(lèi)型分別構(gòu)建了一種互聯(lián)網(wǎng)電視節(jié)目流行度預(yù)測(cè)模型BD3P，并將BD3P模型應(yīng)用于實(shí)際數(shù)據(jù)測(cè)驗(yàn)。實(shí)驗(yàn)結(jié)果表明，與現(xiàn)有其他方法相比，BD3P模型預(yù)測(cè)精度可提升17%以上，并能有效縮短預(yù)測(cè)周期。
-
關(guān)鍵詞:
- 互聯(lián)網(wǎng)電視 /
- 流行度預(yù)測(cè) /
- 行為動(dòng)力學(xué) /
- 最小二乘支持向量機(jī) /
- 雙種群粒子群優(yōu)化
Abstract: Predicting program popularity is a key issue for design and optimization of Internet TV system. Existing prediction methods usually need large quantity of samples and long training time, while the prediction accuracy is poor for the burst hot programs. This paper introduces an Internet TV Program Popularity Prediction model based on viewing Behavioral Dynamics features (BD3P). 6 billion view behavior records from 2.8 million subscribers of a certain Internet TV platform are measured, and the evolution process of program popularity is divided into 4 types based on behavioral dynamics features, which is endogenous, internal subcritical, exogenous and exogenous subcritical. The prediction models of Internet TV program popularity are constructed for each type using Least Squares Support Vector Machines (LSSVM) with double population Particle Swarm Optimization (PSO), and these models are applied to the actual data test. The experimental results show that, compared to the existing prediction model, the prediction accuracy can be increased by more than 17%, and the forecast period can be effectively shortened. -
朱軼, 糜正琨, 王文鼐. 一種基于內(nèi)容流行度的內(nèi)容中心網(wǎng)絡(luò)緩存概率置換策略[J]. 電子與信息學(xué)報(bào), 2013, 35(6): 1305-1310. doi: 10.3724/SP.J.1146.2012.01143. ZHU Yi, MI Zhengkun, and WANG Wennai. A cache probability replacement policy based on content popularity in content centric networks[J]. Journal of Electronics Information Technology, 2013, 35(6): 1305-1310. doi: 10.3724 /SP.J.1146.2012.01143. 芮蘭蘭, 彭昊, 黃豪球, 等. 基于內(nèi)容流行度和節(jié)點(diǎn)中心度匹配的信息中心網(wǎng)絡(luò)緩存策略[J]. 電子與信息學(xué)報(bào), 2016, 38(2): 325-331. doi: 10.11999/JEIT150626. RUI Lanlan, PENG Hao, HUANG Haoqiu, et al. Popularity and centrality based selective caching scheme for information- centric networks[J]. Journal of Electronics Information Technology, 2016, 38(2): 325-331. doi: 10.11999/JEIT150626. GMEZ V, KALTENBRUNNER A, and LPEZ V. Statistical analysis of the social network and discussion threads in slashdot[C]. ACM International Conference on World Wide Web, Beijing, China, 2008: 645-654. doi: 10.1145 /1367497.1367585. SZABO G and HUBERMAN B A. Predicting the popularity of online content[J]. Communications of the ACM, 2010, 53(8): 80-88. doi: 10.1145/1787234.1787254. CASTILLO C, ELHADDAD M, PFEFFER J, et al. Characterizing the life cycle of online news stories using social media reactions[C]. ACM International Conference on Computer Supported Cooperative Work Social Computing, Baltimore, MD, USA, 2014: 211-223. doi: 10.1145/2531602. 2531623. PINTO H, ALMEIDA J M, and GONALVES M A. Using early view patterns to predict the popularity of YouTube videos[C]. ACM International Conference on Web Search and Data Mining, Rome, Italy, 2013: 365-374. doi: 10.1145/ 2433396.2433443. GAO S, MA J, and CHEN Z. Modeling and predicting retweeting dynamics on microblogging platforms[C]. ACM International Conference on Web Search and Data Mining, Shanghai, China, 2015: 107-116. doi: 10.1145/2684822. 2685303. CRANE R and SORNETTE D. Robust dynamic classes revealed by measuring the response function of a social system[J]. Proceedings of the National Academy of Sciences of the United States of America, 2008, 105(41): 15649-15653. doi: 10.1073/pnas.0803685105. WU B, MEI T, CHENG W H, et al. Unfolding temporal dynamics: Predicting social media popularity using multi-scale temporal decomposition[C]. Thirtieth AAAI Conference on Artificial Intelligence, Phoenix, AZ, USA, 2016: 32-38. doi: 10.13140/RG.2.2.27504.66565. WU J, ZHOU Y, CHIU D M, et al. Modeling dynamics of online video popularity[C]. IEEE International Symposium on Quality of Service, Portland, OR, USA, 2015: 141-146. doi: 10.1109/IWQoS.2015.7404724. FONTANINI G, BERTINI M, and DEL BIMBO A. Web video popularity prediction using sentiment and content visual features[C]. ACM International Conference on Multimedia Retrieval, New York, NY, USA, 2016: 289-292. doi: 10.1145/2911996.2912053. ZAMAN T, FOX E B, and BRADLOW E T. A Bayesian approach for predicting the popularity of tweets[J]. The Annals of Applied Statistics, 2014, 8(3): 1583-1611. doi: 10.1214/14-AOAS741. WANG J, ZHANG Z, and ZHANG W. Support vector machine based on double-population particle swarm optimization[J]. Journal of Convergence Information Technology, 2013, 8(8): 33-43. doi: 10.4156/jcit.vol8.issue8. 106. -
計(jì)量
- 文章訪問(wèn)數(shù): 1447
- HTML全文瀏覽量: 169
- PDF下載量: 252
- 被引次數(shù): 0
下載:
