The Research Of The Sequence Mining Algorithms On The High-Dimension Time Series Based On Human Motion Capture Data

Every time series can be considered as a time-varying sequence, which exists in many areas such as finance, economy, natural science and information engineering. Researches on how to process and mine these time series have been paid great attentions and many classic problems have been solved efficiently. In recent years, as both the technologies and the requirements developed rapidly, many large-scale and complex time series databases have been established. Unfortunately, because of the highly increasing computation complexity, most of the data mining technologies, which can be applied on the one-dimension time series successfully, are not available when used to deal with high-dimension time series.Aiming at this problem, we have made a systematical summarization on the researches of data mining on time series and a comprehensive analysis on the characteristics, varieties and current research situation of the high-dimension time series. Based on the review the existing technologies, we focus on two popular topics: the fast sequence searching and motif mining on high-dimension time series. By taking the human motion capture data as the analysis object, effective algorithm is brought forward on the two subjects respectively and the experiments verify the availability and high performance.(1) Different from the spatial position based model used in traditional retrieval methods, a novel model based on kinetic energy is proposed and used to describe human motion. Based on the model, the motion coordination is introduced and used to extract the low-dimensional indexing sequences that are remarkable features of original motions. The support vector machine is applied to classify the motion. Finally, the sequential indexing algorithm based on the Keogh lower bound is employed to measure the similarity of query motion and candidate motions exactly.(2) Against the problem that existing algorithms are apt to be interfered by noise, a motif mining algorithm based on the LCSS distance is introduced. The algorithm has pruned efficiently by using the heuristic strategy based on the distance between subsequences during the search. Then the MDL principle is used to calculate the weights of the unequal-length candidate sequences based on which the motif patterns are selected.