Martha Dais Ferreira's repositories
AISclassification
A geometric-driven semi-supervised approach for fishing activity detection from AIS data.
TrajectoriesCompressionAnalysis
The proposed methodology assess how compression algorithms influence the clustering analysis with respect to anomaly detection of vessel trajectories.
MovementAnalysis
An analysis of the behavior of fishing vessels in order to detect patterns and outliers/anomalies.
NeuralNetworkTrainingFingerprint
Neural Network Training Fingerprint (NNTF) is a visualization approach to analyze the training process of any neural network performing classification.
shaplets-python
Python implementation of the Learning Time-Series Shapelets method, that learns a shapelet-based time-series classifier with gradient descent.
studyForwardCNN
This source codes are used to execute the forward process of Convolutional Neural Networks (CNN) for images input. Such source code was written in the C++ Language, in which can be running in sequential or in parallel. In addition, a source code of Multilayer Perceptron (MLP) was developed in R language to study the impact of dropout and regularization terms during the learning process.
Estimate-Parameters
This source codes are used to estimate the most adequate convolution mask sizes and number of units for a convolution layer of CNN. Based on the False Nearest Neighbors (FNN), a well-known tool from the area of Dynamical Systems, this method helps estimating CNN architectures that are both less complex and more effective.
MSongsDB
Code for the Million Song Dataset, the dataset contains metadata and audio analysis for a million tracks, a collaboration between The Echo Nest and LabROSA. See website for details.