From: Context-aware rule learning from smartphone data: survey, challenges and future directions
Base technique | Description | References |
---|---|---|
Single parameter | A predefined value of time interval, e.g., 15 min is used to generate segments | Ozer et al. [60] |
A different value of time interval, e.g., 30 min is used for segmentation | ||
A relatively large value of the parameter, e.g., 2-h is used to generate time segments | Karatzoglou et al. [63] | |
Another large value of time interval, e.g., 3-h is used for segmentation to make the number of segments small | Phithakkitnukoon et al. [64] | |
Calendar | Various calendar schedules and corresponding time boundaries are used to model users’ behavior in temporal context | Khail et al. [65], Dekel et al. [66], Zulkernain et al. [67], Seo et al. [68], Sarker et al. [28, 59] |
Multi-thresholds | To identify the lower and upper boundary of a particular segment for the purpose of segmenting time-series log data | Halvey et al. [38] |
Data shape | A data shape based time-series data analysis | |
Sliding window | A sliding window is used to analyze time-series data | |
Clustering | A predefined number of clusters is used to discover rules from time-series data | Das et al. [72] |
Genetic algorithm | A genetic algorithm is used to analyze time-series data |