China 
Africa 
Explore chapters and articles related to this topic
Analysis of Human Walking Trajectories for Surveillance
Published in Yangsheng Xu, Ka Keung C. Lee, Human Behavior Learning and Transfer, 2005
The advantage of LCSS is that it allows some of the elements in the trajectories to be unmatched, which can avoid the problem of outliers as in the cases of Euclidean distance and DTW. Also, the algorithm involved in LCSS is computationally efficient, especially compared to the computation required to determine the Lp norm in DTW and the iterative algorithm of HMM. However, the longest common subsequence approach also has its own limitation. It has difficulties in differentiating the sequences that have the same length of longest common subsequences but different gap sizes in between. In our implementation, the “gaps” in the trajectories occur when a person walks away from the expected normal path for a short while and then returns to the normal path. This short-term abnormality can be picked up by the SVM-based trajectory point normality classifier. Therefore the limitation of LCSS does not affect the performance of our overall performance to a significant extent.
Defending Web Applications Against JavaScript Worms on Core Network of Cloud Platforms
Published in Brij B. Gupta, Michael Sheng, Machine Learning for Computer and Cyber Security, 2019
Shashank Tripathi, Pranav Saxena, Harsh D. Dwivedi, Shashank Gupta
Template Builder: This module is responsible for creating script templates. The inputs to this module are script congregates. The module finds Longest Common Subsequence (LCS) of the scripts of the given congregate. The module then finds a script which is most similar to the LCS (its Levenshtein’s distance with LCS is minimum), let it be string s. The module then compares string s and LCS. Positions at which character matches, the character at that position is appended in the template string. Else, ‘-’ character is appended in the template string, only if the previous character in template string is not ‘-’ character. Figure 8 highlights an example of template building of a given congregate.
Author’s Solutions to PL-EN Corpora Processing Problems
Published in Krzysztof Wołk, Machine Learning in Translation Corpora Processing, 2019
However, none were found in our corpus; therefore, the experiments were constrained to small clusters with two pairs of sentences. Matching sentences from the parallel corpus were identified in every cluster. This allowed the generation of new, similar sentences, which were present in the training corpus. For each of the sequential analogies that were identified, a rewriting model is constructed. This was achieved by string manipulation. Common prefixes and suffixes for each of the sentences were calculated using the Longest Common Subsequence (LCS) method [196].
Enhancing requirements reusability through semantic modeling and data mining techniques
Published in Enterprise Information Systems, 2018
Themistoklis Diamantopoulos, Andreas Symeonidis
We use the Longest Common Subsequence (LCS) (Cormen et al. 2009) in order to determine the similarity score between two sequences. The LCS of two sequences and is defined as the longest subsequence of common (yet not consecutive) elements between them. Given, e.g., the sequences and , their LCS is . Finally, the similarity score between the sequences is defined as: