API Reference

This is the main reference for classes and functions present in trajminer.

trajminer: Base

Library essential classes and resources.

Classes

trajminer.TrajectoryData(attributes, data, tids)

Trajectory data wrapper.

trajminer.classification: Classification

Trajectory classifiers. Implementations of traditional classification algorithms work alongside with a trajectory similarity measure from trajminer.similarity.

Classes

classification.KNearestNeighbors([…])	K-Nearest Neighbors Classifier.
classification.TraClass()	TraClass: Trajectory Classification Using Hierarchical Region-Based and Trajectory-Based Clustering.

trajminer.clustering: Clustering

Clustering algorithms for trajectories. Implementations of traditional clustering algorithms work alongside with a trajectory similarity measure from trajminer.similarity.

Classes

clustering.AgglomerativeClustering(n_clusters)	Hierarchical Agglomerative Clustering.
clustering.DBSCAN([eps, min_samples, …])	DBSCAN Clustering.
clustering.KMedoids(n_clusters[, init, …])	K-Medoids Clustering.

trajminer.similarity: Similarity

Trajectory similarity measures.

Classes

similarity.EDR(dist_functions, thresholds)	Edit Distance on Real sequence.
similarity.LCSS(dist_functions, thresholds)	Longest Common SubSequence.
similarity.MSM(dist_functions, thresholds, …)	Multidimensional Similarity Measure.
similarity.MUITAS(dist_functions, …)	Multiple-Aspect Trajectory Similarity Measure.

Functions

similarity.pairwise_similarity(X[, Y, …])

Computes the similarity between trajectories in X and Y.

trajminer.utils: Utils

Utilities for supporting other classes and resources of the library.

Classes

utils.CSVTrajectoryLoader(file[, sep, …])

A trajectory data loader from a CSV file.

Functions

utils.distance.discrete(x, y)	Computes the discrete distance between two objects.
utils.distance.euclidean(x, y)	Computes the euclidean distance between two objects described by an array of floats.
utils.distance.haversine(x, y[, unit])	Computes the haversine distance between two pairs of latitude and longitude.

trajminer.preprocessing: Preprocessing

Preprocessing tools.

Classes

preprocessing.OneHotEncoder()	Encode trajectory features as the concatenation of the one-hot numeric array for each feature.
preprocessing.TrajectorySegmenter(attributes)	Trajectory segmenter.

Functions

preprocessing.filter_trajectory_length(data, …)	Removes trajectories by length criteria.
preprocessing.filter_label_size(data, …[, …])	Removes trajectories corresponding to sets of labels that do not meet size criteria.
preprocessing.filter_duplicate_points(data, …)	Removes duplicates of trajectory points according to the given criteria.