臺灣博碩士論文加值系統

The rapid growth of smartphone uses in today’s modern life encourages the development of useful application that provides a batch of useful information to it's users. The main aim of this development is to make smartphone smarter. Phone call prediction is one of applications that serve as an important feature to achieve smarter smartphone. To the best of our knowledge, there were two most recent works on the development of telephone call [11, 12]. Even though those studies have shown the promising achievements over the basic congenital telephone system, we are still confident to explore many basic features that seem that received little attention in their research. In this paper, we investigate more conservative features that can subscribe as same accuracy result or even better. More specifically, given the user historical call activities, we explore four major features: frequency, duration, recency, and direction. The frequency feature refers the number of interaction calls between the user and callee. The period of time that the user and callee spent in each of their communication call defines as duration feature. While, recency feature is the weight of each connection call between user and callee according to the recentness of that call. Lastly, the feature of direction describes the importance of call initiator between user and callee by giving pre-defined weight. According to these features, we develop three probability ranking models: Probability General-Frequency (PGF), Probability General-Duration (PGD), and Probability Recency (PR). Moreover, we train these models in two real Call Detail Record (CDR) datasets, Reality Mining and Chunghua Telecom dataset to gain depiction result. Finally, we compare these models with the existing works and demonstrate that our conventional models can reach same and even better accuracy prediction.

The rapid growth of smartphone uses in today’s modern life encourages the development of useful application that provides a batch of useful information to it's users. The main aim of this development is to make smartphone smarter. Phone call prediction is one of applications that serve as an important feature to achieve smarter smartphone. To the best of our knowledge, there were two most recent works on the development of telephone call [11, 12]. Even though those studies have shown the promising achievements over the basic congenital telephone system, we are still confident to explore many basic features that seem that received little attention in their research. In this paper, we investigate more conservative features that can subscribe as same accuracy result or even better. More specifically, given the user historical call activities, we explore four major features: frequency, duration, recency, and direction. The frequency feature refers the number of interaction calls between the user and callee. The period of time that the user and callee spent in each of their communication call defines as duration feature. While, recency feature is the weight of each connection call between user and callee according to the recentness of that call. Lastly, the feature of direction describes the importance of call initiator between user and callee by giving pre-defined weight. According to these features, we develop three probability ranking models: Probability General-Frequency (PGF), Probability General-Duration (PGD), and Probability Recency (PR). Moreover, we train these models in two real Call Detail Record (CDR) datasets, Reality Mining and Chunghua Telecom dataset to gain depiction result. Finally, we compare these models with the existing works and demonstrate that our conventional models can reach same and even better accuracy prediction.

1 Introduction 1
2 Related Work 4
2.1 Mobile device log analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.2 Telephone Call Forecasting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.3 Call Prediction Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3 Dynamic Phonebook Framework 8
3.1 Framework Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.2 Preliminaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.3 Dataset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.3.1 Chunghwa Dataset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.3.2 Reality Mining Dataset (RMD) . . . . . . . . . . . . . . . . . . . . . . 11
4 Method 13
4.1 Feature Discovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
4.1.1 Frequency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
4.1.2 Duration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4.1.3 Recency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4.1.4 Direction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4.2 Call Predictor Based Probability Model . . . . . . . . . . . . . . . . . . . . . . 15
4.2.1 General-Frequency Probability . . . . . . . . . . . . . . . . . . . . . . . 15
4.2.2 General-Duration Probability . . . . . . . . . . . . . . . . . . . . . . . 15
4.2.3 Recency Probability . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
5 Evaluation Result 18
5.1 Evaluation Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
5.2 Competitor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
5.3 Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
5.4 Evaluation Result . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
5.4.1 Prediction Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
5.4.2 Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
5.4.3 Time Computation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
6 Conclusion

[1] L. Akoglu and B. Dalvi. Structure, tie persistence and event detection in large phone
and sms networks. In Proceedings of the Eighth Workshop on Mining and Learning with
Graphs, pages 10–17. ACM, 2010.
[2] B. H. Andrews and S. M. Cunningham. Ll bean improves call-center forecasting. Inter-
faces, 25(6):1–13, 1995.
[3] L. Bianchi, J. Jarrett, and R. Choudary Hanumara. Improving forecasting for telemarketing
centers by arima modeling with intervention. International Journal of Forecasting,
14(4):497–504, 1998.
[4] L. Bianchi, J. E. Jarrett, and R. Choudary Hanumara. Forecasting incoming calls to
telemarketing centers. Journal of Business Forecasting Methods and Systems, 12:3–3,
1993.
[5] J. M. Boulin. Call center demand forecasting: improving sales calls prediction accu-
racy through the combination of statistical methods and judgmental forecast. PhD thesis,
Massachusetts Institute of Technology, 2010.
[6] V. R. Carvalho and W. W. Cohen. Ranking users for intelligent message addressing. In
Advances in Information Retrieval, pages 321–333. Springer, 2008.
[7] D. Choujaa and N. Dulay. Predicting human behaviour from selected mobile phone data
points. In Proceedings of the 12th ACM international conference on Ubiquitous computing,
pages 105–108. ACM, 2010.
[8] N. Eagle, A. S. Pentland, and D. Lazer. Mobile phone data for inferring social network
structure. In Social computing, behavioral modeling, and prediction, pages 79–88.
Springer, 2008.
[9] J. Froehlich, M. Y. Chen, S. Consolvo, B. Harrison, and J. A. Landay. Myexperience: a
system for in situ tracing and capturing of user feedback on mobile phones. In Proceedings
of the 5th international conference on Mobile systems, applications and services, pages
57–70. ACM, 2007.
[10] C. E. Harless and T. J. Kowalski. System and method for correlating incoming and
outgoing telephone calls using predictive logic, July 4 2000. US Patent 6,084,954.
[11] S. Phithakkitnukoon and R. Dantu. Towards ubiquitous computing with call prediction.
ACM SIGMOBILE Mobile Computing and Communications Review, 15(1):52–64, 2011.
[12] S. Phithakkitnukoon, R. Dantu, R. Claxton, and N. Eagle. Behavior-based adaptive call
predictor. ACM Transactions on Autonomous and Adaptive Systems (TAAS), 6(3):21,
2011.
[13] M. Roth, A. Ben-David, D. Deutscher, G. Flysher, I. Horn, A. Leichtberg, N. Leiser,
Y. Matias, and R. Merom. Suggesting friends using the implicit social graph. In Pro-
ceedings of the 16th ACM SIGKDD international conference on Knowledge discovery and
data mining, pages 233–242. ACM, 2010.
[14] N. Singh, S. Bagchi, and Y.-S. Wu. Annoying telephone-call prediction and prevention,
May 4 2009. US Patent App. 12/434,750.
[15] V. Soto, V. Frias-Martinez, J. Virseda, and E. Frias-Martinez. Prediction of socioeconomic
levels using cell phone records. In User Modeling, Adaption and Personalization,
pages 377–388. Springer, 2011.
[16] D. Zhang, A. V. Vasilakos, and H. Xiong. Predicting location using mobile phone calls.
ACM SIGCOMM Computer Communication Review, 42(4):295–296, 2012.
[17] H. Zhang and R. Dantu. Predicting social ties in mobile phone networks. In Intelligence
and Security Informatics (ISI), 2010 IEEE International Conference on, pages 25–30.
IEEE, 2010.