Imputing missing values using Inverse Distance Weighted Interpolation for time series data

Data mining is the process of analyzing and retrieving meaningful information from a database. Temporal data mining deals with time stamped data. In the real world, temporal data obtained may contain noisy, inconsistent data and in most cases the data may be missing; hence data preprocessing is one of the important steps that has to be carried out in data mining. Missing values may generate biased results and affect the accuracy of classification. In order to overcome this it is necessary to impute the missing values based on other information in the dataset. The work focuses on imputing missing values using Inverse Distance Weighted Interpolation method which best suits for data sampled at uneven intervals of time. This method assigns values to unknown points from a weighted sum of values of known points. Machine learning techniques applied to the imputed dataset will give better accuracy than that of the incomplete dataset.