Similar Sequential Data Search Algorithm Based on Dimension-By-Dimension Strategy
Because the traditional similar sequential data search algorithm considers only one-dimensional data, its data search accuracy is low, and the search data is not comprehensive. Hence, a similar sequential data search algorithm based on the dimension-by-dimension strategy is proposed. The algorithm measures the similar time series data in order to fill in the missing data in a time series, searches similar sub-sequences in time series data based on the strategy of dimension-by-dimension according to the data measurement results, gives a similarity threshold, queries the dynamic time-bending distance between sequences and the starting position of sub-sequences and candidate sub-sequences according to the threshold, and obtains the data by dynamically adjusting the search target hierarchical matching and search tasks. The experimental results show that under the influence of different levels of interference data, compared with the traditional search algorithm, the search matching accuracy of the proposed search algorithm is maintained at a relatively high level, the data search can obtain 89% of the expected search amount. Moreover, the process takes less time, demonstrating that the performance of the algorithm is superior, and can meet the current search requirements of similar time series data.
Keywords: Dimension-by-dimension strategy, similar time series data, search algorithm, similarity threshold.
Ying Liu, H. Pan, "Similar Sequential Data Search Algorithm Based on Dimension-By-Dimension Strategy",
Engineering Intelligent Systems, vol. 30 no. 3, pp. 233-242, 2022.