A quality evaluation model for Android system based on forum text mining
Cangzhou Yuan, Yue You, Shenhong Wei, Naiyu Yin · 2016
The popularity of Android intelligent terminal has increased the need to evaluate Android operating system. However, traditional evaluation models largely aim at common software and overlook the characteristics of Android. Moreover, the selection of evaluation index is limited to expert's experience and questionnaire surveys. In this paper, we develop a text mining based model to evaluate Android system quality scientifically. We collected massive data from users' comments of mainstream Android forums and Android system evaluation reports. Further extracted evaluation index from users' comments through text mining methods and obtained index weight by counting the frequency of the index. Based on ISO/IEC9126, we classified the evaluation index and designed a multi tree based algorithm—Meval to calculate the evaluation result. Finally, we made a comparative analysis with the traditional evaluation model via huge actual test cases. Results show that our model can not only ensured the accuracy of the evaluation, but also optimize itself with the changing of users' quality needs.