Novel imputation for time series data

Time series data are used in a large variety of real-world applications. However, they often encounter the missing value problem due to data transmission errors, machine malfunction, or human errors. Existing imputation methods for missing values don't explicitly employ the temporal information embedded in the time series data. In this paper, we propose a new imputation method to fill up the missing slots in the time series data. A local time index scheme is developed for taking advantage of the temporal information. By presenting explicitly local time indices and non-missing values to the least squares support vector machine (LSSVM), missing values can be computed easily and naturally. The effectiveness of our proposed method is demonstrated by experiments done on real-world time series datasets.