﻿ 网络电视用户行为计算建模与统计分析 Calculation Modeling and Statistical Analysis of Network TV User Behavior

Computer Science and Application
Vol. 09  No. 01 ( 2019 ), Article ID: 28688 , 9 pages
10.12677/CSA.2019.91021

Calculation Modeling and Statistical Analysis of Network TV User Behavior

Songtao Wu1, Yi Wang2, Lan Wang3, Can Yang3*

1Guangzhou Digital Media Group Co., Ltd., Guangzhou Guangdong

2Tencent, Shenzhen Guangdong

3School of Computer Science & Engineering, South China University of Technology, Guangzhou Guangdong

Received: Jan. 7th, 2019; accepted: Jan. 22nd, 2019; published: Jan. 29th, 2019

ABSTRACT

Network television (IPTV) services are becoming more and more popular in China, and in-depth exploration of audience behavior has become an emerging topic in IPTV. We define a state model to understand network TV users’ behaviors, and conduct users’ behaviors’ computation and analysis from such 3 aspects as an individual user, groups and a single channel in a large-scale IPTV system. We also focus on statistical analysis sampling from the watching length, the number of the online and the distribution of 24 hours, the result of which will be provided some references for delivering web content and advertisement, resources allocation of television channels and the precise content recommendation.

Keywords:IPTV, Live Streaming, User Behavior, Statistics Analysis

1广州珠江数码集团股份有限公司，广东 广州

2深圳市腾讯计算机系统有限公司，广东 深圳

3华南理工大学计算机科学与工程学院，广东 广州

1. 介绍

2. 定义与模型

2.1. 数据集与数据结构

2.2. 观看行为建模

Table 1. IPTV audience user behavior characteristics basic parameter definition table

S：Surfing浏览状态，该状态下的用户表现为快速浏览多个电视频道，一般采用顺序方法依次在相邻频道间进行切换，发现其观看意愿，其特征为每个被浏览的频道的停留时间极短。记作： $\left\{S|01,t ，这里ε表示一个大于零的足够小的观看时长。

V：Viewing欣赏状态，该状态下的用户表现为停止切换，停留在一个频道上观看，不发生频道跳转，且观看时间较长。记作： $\left\{V|d>\epsilon ,C=1,\epsilon

I：Idle状态，该状态下用户表现为既不继续观看某一频道，又不进行频道切换，系统处于空闲状态。为了保障研究的一致性，本文定义一个空频道Null Channel，若用户处于空频道(频道号为0，ch0)，即用户处于空闲期，没有观看任何频道内容。记作： $\left\{I|d=0,C=0,0

① 关闭close，包括两个子行为V → I，S → I；

② 开启start，包括两个子行为I → V，I → S；

③ 换台change，包括两个子行为V → V，S → V；

④ 浏览surfing，包括两个子行为V → S，S → S。

Figure 1. Network TV user viewing behavior modeling state diagram

$\left\{I\left(tt8\right)\right\}$

Figure 2. Schematic diagram of the user-based viewing process based on the timeline

(a) (b)(c) (d)

Figure 3. User watch time CDF chart. (a) x watch duration, y probability distribution; (b) pair (a) x-axis logarithm; (c) partition enlargement x-axis; (d) Meeyoung’ Result

3. 网络电视数据统计分析

3.1. 群体行为的统计分析

(a) (b)(c) (d)

Figure 4. Online population statistics. (a) The number of online users of the entire IPTV user group; (b) Number of online users of ch1; (c) Number of online users of ch7; (d) Number of online users of ch26

3.2. 典型频道统计分析

3.3. 个体用户统计分析

(a) 平均观看率 (b) 平均逃逸率 (c) 在线人数 (d) 在线人数变化率

Figure 5. 24-hour online population statistics for ch1

(a) 用户1 (b) 用户2

Figure 6. User 1 and User 2 time distribution of one month

(a) (b)

Figure 7. Time distribution of the number of views of user1time view and user 2 favorite channel. (a) User 1; (b) User 2

4. 结束语

Calculation Modeling and Statistical Analysis of Network TV User Behavior[J]. 计算机科学与应用, 2019, 09(01): 172-180. https://doi.org/10.12677/CSA.2019.91021

1. 1. Qiu, T., Ge, Z., Lee, S., et al. (2009) Modeling User Activities in a Large IPTV System. Proceedings of the 9th ACM SIGCOMM Con-ference on Internet Measurement, Chicago, Illinois, 4-6 November 2009, 430-441. https://doi.org/10.1145/1644893.1644945

2. 2. Abrahamsson, H. and Nordmark, M. (2012) Program Popularity and Viewer Behavior in a Large TV-on-Demand System. Proceedings of the 2012 Internet Measurement Conference, Boston, Massachusetts, 14-16 November 2012, 199-210.

3. 3. Wang, B., Wen, X.M., Sun, Y., et al. (2009) A New Approach Measuring Users’ QoE in the IPTV. IEEE 2009 Pacific-Asia Conference on Circuits, Communications and Systems, Chengdu, 16-17 May 2009, 453-456. https://doi.org/10.1109/PACCS.2009.63

4. 4. Cha, M., Rodriguez, P., Crowcroft, J., et al. (2008) Watching Television over an IP Network. Proceedings of the 8th ACM SIGCOMM Conference on Internet Measurement, Vouliagmeni, Greece, 20-22 October 2008, 71-83, 71-84. https://doi.org/10.1145/1452520.1452529

5. 5. Yang, C. and Liu, Y. (2015) On Achieving Short Channel Switching Delay and Playback Lag in IP-Based TV Systems. IEEE Transactions on Multimedia, 17, 1096-1106. https://doi.org/10.1109/TMM.2015.2429552

6. 6. Yang, C., Ren, S.X., Liu, Y., Cao, H.W., Yuan, Q.H. and Han, G.Q. (2018) Personalized Channel Recommendation Deep Learning from a Switch Sequence. IEEE Access, 6, 50824-50838. https://doi.org/10.1109/ACCESS.2018.2869470

7. NOTES

*通讯作者。