Smartphone-based activity recognition (SP-AR) recognizes users’ activities using the embedded accelerometer sensor. Only a small number of previous works can be classified as online systems, i.e., the whole process (pre-processing, feature extraction, and classification) is performed on the device. Most of these online systems use either a high sampling rate (SR) or long data-window (DW) to achieve high accuracy, resulting in short battery life or delayed system response, respectively. This paper introduces a real-time/online SP-AR system that solves this problem. Exploratory data analysis was performed on acceleration signals of 6 activities, collected from 30 subjects, to show that these signals are generated by an autoregressive (AR) process, and an accurate AR-model in this case can be built using a low SR (20 Hz) and a small DW (3 s). The high within class variance resulting from placing the phone at different positions was reduced using kernel discriminant analysis to achieve position-independent recognition. Neural networks were used as classifiers. Unlike previous works, true subject-independent evaluation was performed, where 10 new subjects evaluated the system at their homes for 1 week. The results show that our features outperformed three commonly used features by 40% in terms of accuracy for the given SR and DW.

Context: Self-adaptive systems have been studied in software engineering over the past few decades attempting to address challenges within the field. There is a continuous significant need to fully understand the behavior and characteristics of the systems that operate in dynamic environments. By learning the behavior pattern of the environment, we can avoid unnecessary adaptations imbalance efforts for adaptation. As such, there exist research in the area of machine learning aimed at understanding dynamic environments regarding self-adaptive systems. Objective: This study aims to help software practitioners to address adaptation concerns by performing a systematic literature review that provides a comprehensive overview of using machine learning (ML) in self-adaptive systems. We summarize state-of-the-art Of the ML approaches used to handle self-adaptation to help software engineers in the proper selection of ML techniques based on the adaptation concern. Method: This review examines research published between 2001 and 2019 on ML implementation in self-adaptive systems, focusing on the adaptation aspects and purposes. The review was conducted by analyzing major scientific databases that resulted in 78 primary studies from 315 papers from an automatic search. Result: Finally, this study recommends three future research directions to enhance the application of machine learning in self-adaptive systems.