Preliminary Interpretation Using Ambiguous Queries

Abstract

-Preliminary interpretation supports exploratory search in large databases. The user interacts with it by specifying probability distributions over attributes, which then expresses imprecise conditions about the entities of interest. Preliminary interpretation helps the user on the right query conditions by addressing three key challenges: 1) Efficiently computing results for an imprecise query. 2) Gives out the result of the sensitivity value of individual and combines Queries. 3) Suggesting ideas for respective attributes for the user. Keywords: Interactive data exploration and discovery, Probability Query, Imprecise Queries, Sensitivity Analysis. I. INTRODUCTIONThe main notion of preliminary interpretation using ambiguous queries is helping the user with imprecise queries. The existing system is not able to handle the uncertainty of ambiguous queries, whereas proposed system allows user to express the uncertainty through the probabilistic value. We propose this project, a new approach for exploratory searching large databases. This project provides a novel method to interactively compose imprecise database queries with probabilistic conditions, while providing constant feedback to the user about the most likely results and the potential benefit and risk of each condition. This method is designed to accommodate uncertainty and imprecision in user-provided query conditions through two major technical contributions: A novel notion of sensitivity to quantify the impact of uncertainty on the query result. Fast algorithms for calibrated probability estimation that can adapt to a user-specified real-time constraint on system response time. To illustrate the need for imprecise queries with probabilistic conditions, consider the following example motivated by collaboration with the Cornell Lab of Ornithology. Through hugely successful citizen science projects such as the Lab has collected more than 100 million reports of bird sightings, adding tens of millions annually. It wants to leverage this resource to help less experienced birders identify the species of a bird they observed. Assume each observation in the database specifies properties of the bird (e.g., species, size, color) and the observation event (e.g., location, weather, habitat, and features).