Pattern representation method in time-series data: A survey
R. Dinesh, J. E. Judith · 2025
Internet of Things (IoT) and sensing-related devices are increasing day by day, which produces a large amount of time-series data. Understanding and discovering the knowledge from these data are challenging tasks due to the large volume of data and complexity in terms of storage and processing time. There are different methods in time-series representation that provide a solution to a large volume of data and the velocity of the IoT data stream. In the context of time-series representation, the goal is to decrease the quantity of data points within a time series dataset. It is done in the pre-processing stage in the analytics of IoT-generated time-series data. This research paper reviews and summarizes the previous work that represented the timeseries method in different works. The basics of time-series representation are stated, the similarity and limitations of previous research are reviewed and different possible research areas for future research are determined. We hope that this survey work will help as the stepping stone for those people interested in time series data pattern representation research-related areas.