Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/omidghadami95/geolife-outlierdetection-tmd
Outlier detection (z-score and IQR) and visualization on Geolife dataset for transport mode detection task
https://github.com/omidghadami95/geolife-outlierdetection-tmd
geolife geolife-dataset gis google-colab googlecolab gps iqr keras numpy outlier-detection outlier-detection-datasets pandas smartphone-sensors tensorflow tmd transport-mode-detection visualization z-score
Last synced: 3 days ago
JSON representation
Outlier detection (z-score and IQR) and visualization on Geolife dataset for transport mode detection task
- Host: GitHub
- URL: https://github.com/omidghadami95/geolife-outlierdetection-tmd
- Owner: OmidGhadami95
- Created: 2023-06-23T16:51:45.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-03-08T19:00:06.000Z (10 months ago)
- Last Synced: 2024-03-08T20:04:20.298Z (10 months ago)
- Topics: geolife, geolife-dataset, gis, google-colab, googlecolab, gps, iqr, keras, numpy, outlier-detection, outlier-detection-datasets, pandas, smartphone-sensors, tensorflow, tmd, transport-mode-detection, visualization, z-score
- Language: Jupyter Notebook
- Homepage:
- Size: 349 KB
- Stars: 4
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Geolife-OutlierDetection-visualization-TMD
Outlier detection using IQR and Z-score, and also visualization on Geolife dataset for transport mode detection task.
A comprehensive analysis on Geolife dataset trajectories.After loading and preprocessing the Geolife dataset, we have many CSVs, which include distance, velocity, acceleration, and ... as features (columns). The point is that some records are abnormal. For example, we have some records with more than 100 km/h velocity in the 'walk' label! Hence, it's crucial to remove these incorrect and noisy records. Otherwise, our network will train with inaccurate data, leading to lower accuracy. Moreover, Visualizing data before and after outlier detection is crucial for gaining insights into analyzing the effect of outlier detection. In this code, z-score and IQR have been used for outlier-detection
First, we must detect outliers and then remove all of them. Finally, all records after removing outliers must be merged.# Notice:
1- Use IQR or z-score not both
2- Tune z-score hyperparameter (threshold)
3- This part must perform before calculating aggregated featuresRaw Geolife dataset link : https://download.microsoft.com/download/F/4/8/F4894AA5-FDBC-481E-9285-D5F8C4C4F039/Geolife%20Trajectories%201.3.zip