Trajectory outlier detection based on DBSCAN clustering algorithm

2017 ◽  
Vol 46 (5) ◽  
pp. 528001 ◽  
Author(s):  
周培培 Zhou Peipei ◽  
丁庆海 Ding Qinghai ◽  
罗海波 Luo Haibo ◽  
侯幸林 Hou Xinglin
Keyword(s):  
Outlier Detection ◽  
Clustering Algorithm ◽  
Dbscan Clustering ◽  
Trajectory Outlier Detection

scholarly journals A Quantification Method for Supraharmonic Emissions Based on Outlier Detection Algorithms

Energies ◽  
2021 ◽  
Vol 14 (19) ◽  
pp. 6404
Author(s):  
Hui Zhou ◽  
Zesen Gui ◽  
Jiang Zhang ◽  
Qun Zhou ◽  
Xueshan Liu ◽  
...  
Keyword(s):  
Outlier Detection ◽  
Clustering Algorithm ◽  
Spatial Clustering ◽  
Skewed Distribution ◽  
Distribution Model ◽  
Distribution Data ◽  
Original Spectrum ◽  
Detection Algorithms ◽  
Quantification Method ◽  
Dbscan Clustering

Based on outlier detection algorithms, a feasible quantification method for supraharmonic emission signals is presented. It is designed to tackle the requirements of high-resolution and low data volume simultaneously in the frequency domain. The proposed method was developed from the skewed distribution data model and the self-tuning parameters of density-based spatial clustering of applications with noise (DBSCAN) algorithm. Specifically, the data distribution of the supraharmonic band was analyzed first by the Jarque–Bera test. The threshold was determined based on the distribution model to filter out noise. Subsequently, the DBSCAN clustering algorithm parameters were adjusted automatically, according to the k-dist curve slope variation and the dichotomy parameter seeking algorithm, followed by the clustering. The supraharmonic emission points were analyzed as outliers. Finally, simulated and experimental data were applied to verify the effectiveness of the proposed method. On the basis of the detection results, a spectrum with the same resolution as the original spectrum was obtained. The amount of data declined by more than three orders of magnitude compared to the original spectrum. The presented method will benefit the analysis of quantification for the amplitude and frequency of supraharmonic emissions.

scholarly journals Resolvable Cluster Target Tracking Based on the DBSCAN Clustering Algorithm and Labeled RFS

IEEE Access ◽  
2021 ◽  
Vol 9 ◽  
pp. 43364-43377
Author(s):  
Xirui Xue ◽  
Shucai Huang ◽  
Jiahao Xie ◽  
Jiashun Ma ◽  
Ning Li
Keyword(s):  
Target Tracking ◽  
Clustering Algorithm ◽  
Dbscan Clustering

scholarly journals AN EFFICIENT CLUSTERING METHOD FOR DBSCAN GEOGRAPHIC SPATIO-TEMPORAL LARGE DATA WITH IMPROVED PARAMETER OPTIMIZATION

2020 ◽  
Vol XLII-3/W10 ◽  
pp. 581-584
Author(s):  
J. W. Li ◽  
X. Q. Han ◽  
J. W. Jiang ◽  
Y. Hu ◽  
L. Liu
Keyword(s):  
Parameter Optimization ◽  
Clustering Algorithm ◽  
Optimal Solution ◽  
Large Data ◽  
Parameter Selection ◽  
Physical Analysis ◽  
Clustering Method ◽  
K Value ◽  
Dbscan Clustering ◽  
Spatio Temporal

Abstract. How to establish an effective method of large data analysis of geographic space-time and quickly and accurately find the hidden value behind geographic information has become a current research focus. Researchers have found that clustering analysis methods in data mining field can well mine knowledge and information hidden in complex and massive spatio-temporal data, and density-based clustering is one of the most important clustering methods.However, the traditional DBSCAN clustering algorithm has some drawbacks which are difficult to overcome in parameter selection. For example, the two important parameters of Eps neighborhood and MinPts density need to be set artificially. If the clustering results are reasonable, the more suitable parameters can not be selected according to the guiding principles of parameter setting of traditional DBSCAN clustering algorithm. It can not produce accurate clustering results.To solve the problem of misclassification and density sparsity caused by unreasonable parameter selection in DBSCAN clustering algorithm. In this paper, a DBSCAN-based data efficient density clustering method with improved parameter optimization is proposed. Its evaluation index function (Optimal Distance) is obtained by cycling k-clustering in turn, and the optimal solution is selected. The optimal k-value in k-clustering is used to cluster samples. Through mathematical and physical analysis, we can determine the appropriate parameters of Eps and MinPts. Finally, we can get clustering results by DBSCAN clustering. Experiments show that this method can select parameters reasonably for DBSCAN clustering, which proves the superiority of the method described in this paper.

scholarly journals Analysis of new nosological models from disease similarities using clustering

2020 ◽  
Author(s):  
Lucía Prieto Santamaría ◽  
Eduardo P. García del Valle ◽  
Gerardo Lagunes García ◽  
Massimiliano Zanin ◽  
Alejandro Rodríguez González ◽  
...  
Keyword(s):  
Clustering Algorithm ◽  
Molecular Data ◽  
Modern Medicine ◽  
Biological Information ◽  
Distributed Data ◽  
Feature Vectors ◽  
Dbscan Clustering ◽  
Disease Similarity ◽  
Starting Point ◽  
Evaluation Metric

AbstractWhile classical disease nosology is based on phenotypical characteristics, the increasing availability of biological and molecular data is providing new understanding of diseases and their underlying relationships, that could lead to a more comprehensive paradigm for modern medicine. In the present work, similarities between diseases are used to study the generation of new possible disease nosologic models that include both phenotypical and biological information. To this aim, disease similarity is measured in terms of disease feature vectors, that stood for genes, proteins, metabolic pathways and PPIs in the case of biological similarity, and for symptoms in the case of phenotypical similarity. An improvement in similarity computation is proposed, considering weighted instead of Booleans feature vectors. Unsupervised learning methods were applied to these data, specifically, density-based DBSCAN clustering algorithm. As evaluation metric silhouette coefficient was chosen, even though the number of clusters and the number of outliers were also considered. As a results validation, a comparison with randomly distributed data was performed. Results suggest that weighted biological similarities based on proteins, and computed according to cosine index, may provide a good starting point to rearrange disease taxonomy and nosology.

scholarly journals Lane Formation Beyond Intuition Towards an Automated Characterization of Lanes in Counter-flows

2020 ◽  
Vol 5 ◽  
Author(s):  
Luca Crociani ◽  
Giuseppe Vizzari ◽  
Andrea Gorrini ◽  
Stefania Bandini
Keyword(s):  
Clustering Algorithm ◽  
Variable Density ◽  
Automatic Identification ◽  
Computing Power ◽  
Dbscan Clustering ◽  
Significant Difference ◽  
Lane Formation ◽  
Human Coder ◽  
Behavioural Dynamics

Pedestrian behavioural dynamics have been growingly investigated by means of (semi)automated computing techniques for almost two decades, exploiting advancements on computing power, sensor accuracy and availability, computer vision algorithms. This has led to a unique consensus on the existence of significant difference between unidirectional and bidirectional flows of pedestrians, where the phenomenon of lane formation seems to play a major role. The collective behaviour of lane formation emerges in condition of variable density and due to a self-organisation dynamic, for which pedestrians are induced to walk following preceding persons to avoid and minimize conflictual situations. Although the formation of lanes is a well-known phenomenon in this field of study, there is still a lack of methods offering the possibility to provide an (even semi-) automatic identification and a quantitative characterization. In this context, the paper proposes an unsupervised learning approach for an automatic detection of lanes in multi-directional pedestrian flows, based on the DBSCAN clustering algorithm. The reliability of the approach is evaluated through an inter-rater agreement test between the results achieved by a human coder and by the algorithm.

Sign in / Sign up

Export Citation Format

Share Document