# estimate_state_probs: Estimate conditional/marginal state probabilities In LICORS: Light Cone Reconstruction of States - Predictive State Estimation From Spatio-Temporal Data

## Description

Estimates P(S = s_k; \mathbf{W}), k = 1, …, K, the probability of being in state s_k using the weight matrix \mathbf{W}.

These probabilites can be marginal (P(S = s_k; \mathbf{W})) or conditional (P(S = s_k \mid \ell^{-}, \ell^{+}; \mathbf{W})), depending on the provided information (pdfs$PLC and pdfs$FLC).

• If both are NULL then estimate_state_probs returns a vector of length K with marginal probabilities.

• If either of them is not NULL then it returns an N \times K matrix, where row i is the probability mass function of PLC i being in state s_k, k = 1, …, K.

## Usage

 1 2 estimate_state_probs(weight.matrix = NULL, states = NULL, pdfs = list(FLC = NULL, PLC = NULL), num.states = NULL) 

## Arguments

 weight.matrix N \times K weight matrix states vector of length N with entry i being the label k = 1, …, K of PLC i pdfs a list with estimated pdfs for PLC and/or FLC evaluated at each PLC, i=1, …, N and/or FLC, i=1, …, N num.states number of states in total. If NULL (default) then it sets it to max(states) or ncol(weight.matrix) - depending on which one is provided.

## Value

A vector of length K or a N \times K matrix.

## Examples

 1 2 3 WW <- matrix(runif(10000), ncol = 10) WW <- normalize(WW) estimate_state_probs(WW) 

LICORS documentation built on May 29, 2017, 1:02 p.m.