Trajectory Data Mining: An Overview
Citations
585 citations
547 citations
356 citations
Cites background from "Trajectory Data Mining: An Overview..."
...Index Terms—Big Data, cross-domain datamining, data fusion, multi-modality data representation, deep neural networks, multi-view learning, matrix factorization, probabilistic graphical models, transfer learning, urban computing Ç...
[...]
351 citations
286 citations
Cites background from "Trajectory Data Mining: An Overview..."
...Table 2 demonstrates the taxonomy of VSN applications, which can be further divided into social-data-driven vehicular networks, social vehicular ad hoc networks (VANETs), and data-driven social networks [4]....
[...]
...However, trajectories of vehicles are not perfectly accurate due to sensor noise and other reasons, for example, false positioning signals received in some urban areas [4]....
[...]
References
9,627 citations
"Trajectory Data Mining: An Overview..." refers background in this paper
...The regions whose log-likelihood ratio statistic value drops in the tail of χ2 distribution are likely to be anomalous [Chandola et al. 2009]....
[...]
...A survey on general anomaly detection methods can be found in [14]....
[...]
...The regions whose log-likelihood ratio statistic value drops in the tail of χ(2) distribution are likely to be anomalous [14]....
[...]
3,749 citations
"Trajectory Data Mining: An Overview..." refers methods in this paper
...A well-known algorithm, called Douglas-Peucker [Douglas and Peucker 1973], is used to approximate the original trajectory....
[...]
...The solution first identifies key points shaping a trajectory, by using a line simplification algorithm like DP [Douglas and Peucker 1973]....
[...]
2,082 citations
"Trajectory Data Mining: An Overview..." refers background or methods in this paper
...As the assumption may not hold in reality, Dynamic Time Wrapping (DTW) distance was proposed to allow ‘repeating’ some points as many times as needed in order to get the best alignment [3]....
[...]
...As the assumption may not hold in reality, Dynamic Time Wrapping (DTW) distance was proposed to allow “repeating” some points as many times as needed in order to get the best alignment [Agrawal et al. 1993]....
[...]
...KNN queries retrieve the top K trajectories with the minimum aggregate distance to a few points (entitled the KNN point query [21][94][95]) or a specific trajectory (entitled the KNN trajectory query [117][3])....
[...]
...KNN queries retrieve the top-K trajectories with the minimum aggregate distance to a few points (entitled the KNN point query [Chen et al. 2010; Tao et al. 2002; Tang et al. 2011]) or a specific trajectory (entitled the KNN trajectory query [Yi et al. 1998; Agrawal et al. 1993])....
[...]
1,975 citations
"Trajectory Data Mining: An Overview..." refers methods in this paper
...After the transformation, we can mine the sequential patterns from these sequences by using existing sequential pattern mining algorithms, such as PrefixSpan [Pei et al. 2011] and CloseSpan [Yan et al. 2003], with time constraints....
[...]
...After the transformation, we can mine the sequential patterns from these sequences by using existing sequential pattern mining algorithms, such as PrefixSpan [80] and CloseSpan [112], with time constraints....
[...]
1,903 citations
"Trajectory Data Mining: An Overview..." refers background or methods in this paper
...The noise filtering method, which has been used in T-Drive [Yuan et al. 2010a, 2011a, 2013a] and GeoLife [Zheng et al. 2009a; Zheng et al. 2010] projects, first calculates the travel speed of each point in a trajectory based on the time interval and distance between a point and its successor (we…...
[...]
...The dataset has been used to estimate the similarity between users [Li et al. 2008], which enables friend and location recommendations [Zheng and Xie 2011b; Zheng et al. 2009c]....
[...]
...The dataset has been used to estimate the similarity between users [54], which enables friend and location recommendations [154][155]....
[...]
...[155][154] transform users’ GPS trajectory into a user-location matrix, where a row stands for a user and a column denotes a location (such as a cluster shown in Figure 21)....
[...]
...2011; Zheng et al. 2012b] and travel recommendation [Zheng and Xie 2011b; Zheng et al. 2011c; Zheng et al. 2009b]....
[...]