Trajectory Data Oriented Function Join And Similarity Search Algorithm

Nowadays, the features of information are more explicitly. All the Data got bypeople is stored in database using computer binary information. In order to observeand analyze people’s lives in the physical world better, we use a variety ofinformation and data collection methods. How to do fusion operation to these dataand how to retrieve similar data has become a hot research field. This paper is dividedinto two parts. The first part is about how to use the function join to improve thetraditional method which cannot solve the problem that target trajectory dataassociated with the attribute value data. The second part is about how to do the similarquery with a given target trajectory.In function join algorithm based on target trajectory, the paper first gives thedefinition of the data model of the target data and attribute value data. Thesedefinition reflect the existence of data in a relational database. In this paper, we givethe loop function join algorithm. The algorithm which uses the loop method judgeswhether the join conditions are met by one trajectory and one tips of attribute data.Here the join condition to determine uses function join. The paper use the trajectorydata and the attribute data as the function of the input. The function gives whetherconditions are met through judging whether the data satisfies a similar relationshipbetween time and spatial and whether the attribute data satisfies the attribute valueconstraint. If the condition is satisfied, we say the data joins successfully. Then thepaper gives hash based function join algorithm. This method can improve the loopfunction join algorithm. Hash based method can solve the problem effectively thatloop method cannot use in big data. To improve the two basic algorithm, we proposean optimized strategy. This optimization is to compress the trajectory data, therebyreducing the amount of calculation of join operation. For the optimization strategy,we give the corresponding theoretical analysis and error rate.In similarity search algorithm based on trajectory data, the paper first gives anovel target trajectory distance function metric approach. This approach is aneffective solution to calculate the similarity between the target trajectories, and laysthe foundation for the proposed algorithm later. In this paper, the following definitions and similarity judgment are proposed. Then we use the distance functionof Euclidean space to determine whether the two data are similar. Then we give amethod to determine similarity between trajectories which use a given trajectory datainterception trajectory substring and trajectory data in the database to make similarjudgments algorithms. By this algorithm we are able to query a similar trajectory data.Finally, we present an optimization strategy of similarity search algorithms based onthe target trajectory to improve performance with the recall rate and comparativeexperiment.