Vehicle Tracker fault diagnosis

In this repository is presented some of my work developing some FDI (Fault Detection and Identification) models.

All the code is in Jupyter notebook files, and the data is private.

Each file does the following:

Retirada de dados: Collect de data from a mongoDB instance;
Tratamento de dados vetoriais: handmade feature extration, drastically reducion the dataset dimension;
Análise e preprocessamento - especialista: Pre-processing and data analysis of the new features;
Classificadores B1: Four Classifiers (RF, NB, SVM, MLP) to detect if there is a fault in the tracker;
Classificadores B2: Four Classifiers (RF, NB, SVM, MLP) to detect and identify if there is, and what is the fault in the tracker;
Tratamento para dados crus: Raw data pre-processing, without any manual feature extration;
Dataset split: Spliting the raw dataset (Train/Test split);
CNN 2D - FD: A CNN model to detect faults from raw data (model based on VGG-16);
CNN 2D - FDI: A CNN model to detect and identify faults from raw data (model based on VGG-16)

Name		Name	Last commit message	Last commit date
Latest commit History 12 Commits
Análise e preprocessamento - especialista.ipynb		Análise e preprocessamento - especialista.ipynb
CNN 2D - FD - new.ipynb		CNN 2D - FD - new.ipynb
CNN 2D - FD.ipynb		CNN 2D - FD.ipynb
CNN 2D - FDI.ipynb		CNN 2D - FDI.ipynb
Classificadores B1.ipynb		Classificadores B1.ipynb
Classificadores B2.ipynb		Classificadores B2.ipynb
Dataset split.ipynb		Dataset split.ipynb
README.md		README.md
Retirada de dados.ipynb		Retirada de dados.ipynb
Tratamento de dados vetoriais.ipynb		Tratamento de dados vetoriais.ipynb
Tratamento para dados crus.ipynb		Tratamento para dados crus.ipynb
UNLICENSE.txt		UNLICENSE.txt

Provide feedback