Evaluation of Different Algorithms for Measuring the Similarities of Trajectory Datasets

Abstract—With the techologicaltechnological development of
global positioning system and video survelliancesurveillance
systems and extensive utilization of socailsocial media platforms,
large amounts of trajectories data are captured. Processing and
analysinganalyzing of moving obgectsobjects such as people,
animals, and vehicles provides valuable information for
industrial and academic studies. about them. In this study, three
different similarity algorithms are investigated to find
similaritessimilarities of trajectory data. These algorithms are
Euclidian distance based similarity measurement (ED), Dynamic-
time Warping based similarity measurement (DTW) and Longest
Common Subsequence based similarity measurement (LCSS).
After classifying trajectory data according to their size (small,
medium, large), testsTests are evaluated both on raw data and
reduced data. Reduced data is obtained by using withusing
Douglas-Peucker algorithm. As the data set, our created
synthetic data and Geolife real trajectory dataset are used.The

tests for the evaluation of the similarity algorithms are
performed on both a synthetic dataset and Geolife real trajectory
dataset, with varying data sizes.


Go Here


Büyük Veri, Paralel İşleme ve Akademisyenlik [Link]

Veri Analitiği & Büyük Veri [Link]

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.