智慧型手機在人類的生活中扮演著重要的角色，人們透過手機能夠從事各式各樣的活動，像是：網頁瀏覽、導航、社交等等。越來越多手機上的應用程式是具有高度互動性的，根據人類的互動，即時完成相對應的任務。而我們認為，人們與手機的互動情形會受到使用者情境影響。人們在不同的使用者情境底下，會有各自偏好的習慣，進而影響到手機內的工作負載的產生。我們能夠藉由建立起使用者情境與使用者互動行為的關係，進一步地推測出不同情境底下的手機工作負載。在這篇論文中，我們提出了一個基於使用者情境的工作負載預測模型。我們首先建立使用者情境和使用者行為的關係，透過機器學習的方法將使用者情境進行分類。針對每種情境，我們能夠預測出互動性應用程式的工作負載。我們的實驗結過顯示，我們的預測結果跟實際的CPU使用率平均相差4.21%到6.34%。

Smartphones play an important role in human’s life. People interact with smartphones to perform various activities. More and more mobile applications involve real-time interactions with users. That is, the operation of mobile applications relies on user behavior. However, User behaviors are affected by user contexts. In a certain context, user will perform a fixed behavior preference. Thus, we can infer the user behaviors from user contexts. Moreover, the workloads of the interactive applications can also be estimated. In this paper, we propose a context-based predictive model of smartphone workloads, which aim to build the correlation between user contexts and user behaviors. We apply machine learning methodology to user context classification. By this model, we further predict the workloads of the interactive applications. Our result shows that the average estimated CPU utilization error between predictive result and actual usage is 4.21% to 6.34%.

[1] Seyed Amir Hoseini-Tabatabaei, Alexander Gluhak, and Rahim Tafazolli, “A survey on smartphone-based systems for opportunistic user context recognition”, ACM Computing Surveys (CSUR), vol. 45, no. 3, pp. 27, 2013.
[2] Moshe Unger, Ariel Bar, Bracha Shapira, Lior Rokach, and Ehud Gudes, “Contexto: lessons learned from mobile context inference”, in Proceedings of the 2014 ACM International Joint Conference on Pervasive and Ubiquitous Computing: Adjunct Publication.
ACM, 2014, pp. 175–178.
[3] Hossein Falaki, Ratul Mahajan, Srikanth Kandula, Dimitrios Lymberopoulos, Ramesh Govindan, and Deborah Estrin, “Diversity in smartphone usage”, in Proceedings of the 8th international conference on Mobile systems, applications, and services. ACM, 2010, pp. 179–194.
[4] “Cpu frequency and voltage scaling code in the linux(tm) kernel”, https://www.kernel.org/doc/Documentation/cpu-freq/governors.txt, [Online; accessed 22-August-2016].
[5] Yuhao Zhu, Matthew Halpern, and Vijay Janapa Reddi, “Event-based scheduling for energy-efficient qos (eqos) in mobile web applications”, in 2015 IEEE 21st International Symposium on High Performance Computer Architecture (HPCA). IEEE, 2015, pp. 137–149.
[6] Yongin Kwon, Sangmin Lee, Hayoon Yi, Donghyun Kwon, Seungjun Yang, Byung-Gon Chun, Ling Huang, Petros Maniatis, Mayur Naik, and Yunheung Paek, “Mantis:
automatic performance prediction for smartphone applications”, in Proceedings of the 2013 USENIX conference on Annual Technical Conference. USENIX Association, 2013, pp. 297–308.
[7] Daniel Lo, Taejoon Song, and G Edward Suh, “Prediction-guided performance-energy trade-off for interactive applications”, in Proceedings of the 48th International Symposium on Microarchitecture. ACM, 2015, pp. 508–520.
[8] “Android api guides”, https://developer.android.com/guide/index.html, [Online;
accessed 22-August-2016].
[9] “Google activity recognition service”, https://developers.google.com/android/
reference/com/google/android/gms/location/ActivityRecognition, [Online; accessed 22-August-2016].
[10] “Android accessibilityservice”, https://developer.android.com/reference/
android/accessibilityservice/AccessibilityService.html, [Online; accessed 22-August-2016].
[11] Dongwon Kim, Nohyun Jung, and Hojung Cha, “Content-centric display energy management for mobile devices”, in Proceedings of the 51st Annual Design Automation Conference. ACM, 2014, pp. 1–6.
[12] Haofu Han, Jiadi Yu, Hongzi Zhu, Yingying Chen, Jie Yang, Guangtao Xue, Yanmin Zhu, and Minglu Li, “E 3: energy-efficient engine for frame rate adaptation on smartphones”, in Proceedings of the 11th ACM Conference on Embedded Networked Sensor Systems. ACM, 2013, p. 15.
28