Research On The Efficient Retrieval Strategy In Uncertain Databases

In recent years, uncertain data is widespread and has received extensive attention astechnology advances, data acquisition and processing techniques to understand the deepening ofthe data. In many specific applications, such as military, financial, logistics, GPS positioning,radar, sensor networks of WSN, privacy protection, radio frequency identification RFID field, etc.data uncertainty are widespread.Traditional data retrieval technology has been unable to effectively manage and retrieve data,how to quickly and effectively, to facilitate the analysis of uncertain data in the uncertaindatabases in order to tap potential, valuable and interesting information is becoming more andmore important. At present, according to data of the diversity and application characteristics,experts and scholars proposed a variety of uncertain data model, while the core idea of thesemodels are derived from the possible worlds model. Possible world model is more deterministicdata from a data source of uncertainty evolved, the many data as possible worlds instance, theprobability of all data instances is equal to 1. An uncertainty of data sources can be evolved manyeven exponential instances set of possible worlds, resulting in a number of instances of possibleworlds is far greater than the scale of the uncertain databases. Therefore, we need to takeeffective measures or technologies to deal with less use of information in order to reduceunwanted mass data information, improve query processing efficiency. Uncertainty data miningresearch, based on the possible world model, this paper proposes a method to reduce the largeinstances of possible worlds and reduce the search space. So that we can effectively and quicklyretrieve the interesting information what people want in the information ocean.This paper first describes the research background, domestic and international conditionsand related work. Uncertain data is a newly developed research focus and has received extensiveattention from industry and academia to solve the uncertain data. The challenge has importantsignifance and necessity; Secondly, this paper detailed overview of information retrieval research,including the traditional information retrieval concept, retrieval way, technology and step. Also,including the uncertain data and analysis of its causes. The possible world model is the mostwidely used , the model is the core idea of the uncertainty in data management. The paper alsoraised the uncertain data management problems and challenges faced. Based on the premise ofthe deterministic data, some of the traditional data management theory and technology can not beapplied to management of the uncertain data; And the next chaper introduces three classic queryalgorithm of the uncertain data and gives specific query example analysis. Summarizing of thethree classic query algorithm shortcomings, this paper put forwards the query strategyRPW-kBest to reduce the possible world which could improve the query efficiency and reducethe time cost. Finally, this paper use the association rules in data mining, the knowledge ofconstraint conditions and the R-Tree storage structure to optimize the query algorithm. In thepaper, we also illustrate that the choice of storage structure is a key issue for data search speed,and the space index is the core technology to the increase the retrieval speed. The spatial index isto search for space to provide a suitable data structure to improve the retrieval speed.