Research And Implementation Of Spatial Text Similarity Search

Location based services(LBS) have been widely used in the people’s living and production, for example, mobile Twitter app platform, Google map, etc. With advances in smart phone devices and data technology and popularization, which is associated with a large number of space for text data, smart devices are often equipped with a GPS device. These devices will produce multiple sets of information, which means that this information not only contain keywords, but also contains information about the user’s location. It allows the LBS system a new type number space text information data structures. In the context of big data, for the increasing amount of data systems and how to double for efficient retrieval of information is a very important issue. However, due to the continuity of geographic information, as well as the discrete nature of text messages, and space text message is more difficult in a unified way so that they can combine to be processed, which makes it more challenging.This article did a study on space text keyword search, detailed analysis of the current state of the art spatial query and text query techniques, such as the inverted index, grids, and other methods, and optimization and summarizes some of the methods in recent years. This paper studies the accomplishments and contributions are as follows:First,prefix filtering algorithm for text queries, through our research, existing prefix filtering algorithm, based on a number of optimizations by analyzing the prefix filtering algorithm theory, found prefix filtering threshold upper boundary condition in some similarity calculation on the basis of the original, you can further enhance the filtration efficiency of garbage collection in advance. And by detailed experimental comparison of algorithms, applicability of certain conditions and prove the correctness of the algorithm.Second,spatial indexing method is the most effective way grid and grid methods of grid requires complete and disjoint and uniformity. All grid that covered all the spaces between each pair of grid have nothing in common, and the grid is square, and the same size. This paper intends to design a non-uniform grid of points, space does not necessarily distinguish the uniform grid, thereby reducing the number of spatial grid space text objects in the collection so as to increase the efficiency of queries and filters.In order to overcome the text keyword search, not on the text within the space limitations. A field signature algorithm is presented, primarily adding spatial information in text, arranging space calculated a hash table and text are arranged to form a signature, by some theories to derive similarities and the relationship between the elements of the collection, use of spatial signature of the text space pruning speeds up.