The analysis of time series using data mining techniques can be
effective when all targets have their own inherent patterns in a sparse sampling
acoustic sensor network where no valid feature of frequency can be extracted.
However, both problems of local time shifting and spatial variations should be
solved to deploy the time series analysis. This paper presents time-warped
similarity measure algorithms in order to solve the two problems through time
series, and we propose the IDDC (Improved Derivative DTW-Cosine)
algorithm to deliver the optimal result and prove the performance with some
experiments. The experimental results show that the object classification
accuracy rate of the proposed algorithm outperforms the other time-warped
similarity measure algorithms by at least 10.23%. Since this proposed algorithm
produces such a satisfactory result with sparse sampling data, it allows us to
classify objects with relatively low overhead.