Become an expert in R — Interactive courses, Cheat Sheets, certificates and more!
Get Started for Free

LICORS

Light Cone Reconstruction of States - Predictive State Estimation From Spatio-Temporal Data

Estimates predictive states from spatio-temporal data and consequently can provide provably optimal forecasts. Currently this implementation supports an N-dimensional spatial grid observed over equally spaced time intervals. E.g. a video is a 2D spatial systems observed over time. This package implements mixed LICORS, has plotting tools (for (1+1)D and (2+1)D systems), and methods for optimal forecasting. Due to memory limitations it is recommend to only analyze (1+1)D systems.

Functions (33)

LC-utils

Utilities for LC class

LICORS-package

Light Cone Reconstruction of States - Predictive State Estimation From Spatio-Temporal Data

compute_LC_coordinates

Computes coordinates of PLC and FLC relative to origin

compute_LICORS_loglik

Log-likelihood of LICORS model

compute_NEC

Compute Negative Entropy Criterion (NEC)

compute_margin_coordinates

Get LC configuration from a (N+1)D field

compute_mixture_penalty

Penalty of mixture weights

contCA00

Simulated 7 state (1+1)D field

data2LCs

Iterate over (N+1)D field and get all LC configurations

embed2

Improved embed() function

estimate_LC_pdfs

Estimate PLC/FLC distributions for all states

estimate_state_adj_matrix

Estimate adjacency matrix for equivalent FLC distributions based on states

estimate_state_probs

Estimate conditional/marginal state probabilities

get_LC_config

Get configuration of a light cone (LC)

get_spacetime_grid

Get an iterator over the space-time coordinates of the field.

image2

Improved image() function

initialize_states

State initialization for iterative algorithms (randomly or variants of kmeans)

kmeanspp

Kmeans++

merge_states

Merge several states into one

mixed_LICORS

Mixed LICORS: An EM-like Algorithm for Predictive State Space Estimation

mixed_LICORS-utils

Utilities for “LICORS” class

normalize

Normalize a matrix/vector to sum to one (probability simplex)

predict_FLC_given_PLC

Predict FLCs given new PLCs

rdensity

Generate random sample from density() or wKDE

relabel_vector

Relabels a vector to consecutive labels

remove_small_sample_states

Reassign low sample states to close states

search_knn

K nearest neighbor (KNN) search

setup_LC_geometry

Setup light cone geometry

sparsify_weights

Sparsify weights

states2weight_matrix

Converts label vector to 0/1 mixture weight matrix

threshold

Threshold a matrix/vector below and above

wKDE

Weighted kernel density estimator (wKDE)

weight_matrix2states

Returns unique state assignment from a (row-wise) weight matrix

LICORS

Light Cone Reconstruction of States - Predictive State Estimation From Spatio-Temporal Data

v0.2.0
GPL-2
Authors
Georg M. Goerg <gmg@stat.cmu.edu>
Initial release
2013-11-20

We don't support your browser anymore

Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.