A Novel methodology for Searching Dimension Incomplete Database

This manuscript deals with the similarity querying problems for cases where data loss exists. Limitations in traditional methodologies for querying incomplete data in database, data mining and information retrieval research has urged to shift into development of different new innovative models. This Investigation is done based on a model developed based on ARIMA constructional model to check the performance related to dimensional incomplete data. In actual problems, the data collected by the sensor network in noisy environment results in loss in data values and also dimensional information in some cases. An attempt has been made to verify the effect of time series model on recovering the missing dimension information. The complications encountered while checking out all the possible combinations of missing dimensions has been studied while evaluating the query and the data objects. Implication of the ARIMA model will result in reduction in total time required for searching and recovering the data. It is predicted that time series approach will perform faster that probability triangle model. To reduce the search time, evaluation is done without generating all the recovery versions. Basic concept that sensors give values which have strong resemblance in different times is used in this model. In the time series values, samples at different time will be evaluated to establish consistent ARIMA coefficients. Keywords—Dimension incomplete database, similarity search, ARIMA model.