scan_eb_zip_cpp: Calculate the highest-value EB ZIP loglihood ratio statistic.

In BenjaK/scanstatistics: Space-Time Anomaly Detection using Scan Statistics

Description

Calculate the expectation-based ZIP loglihood ratio statistic for each zone and duration, but only keep the zone and duration with the highest value (the MLC). The estimate of the relative risk is also calculated, along with the number of iterations the EM algorithm performed.

Usage

scan_eb_zip_cpp(counts, baselines, probs, zones, zone_lengths, rel_tol,
  store_everything, num_mcsim)

Arguments

counts	matrix (most recent timepoint in first row)
baselines	matrix (most recent timepoint in first row)
probs	matrix (most recent timepoint in first row)
zones	integer vector (all zones concatenated; locations indexed from 0 and up)
zone_lengths	integer vector
rel_tol	double
store_everything	boolean
num_mcsim	int

Value

A list with elements observed and simulated, each being a data frame with columns:

zone

The top-scoring zone (spatial component of MLC).

duration

The corresponding duration (time-length of MLC).

score

The value of the loglihood ratio statistic (the scan statistic).

relrisk

The estimated relative risk.

n_iter

The number of iterations performed by the EM algorithm.

BenjaK/scanstatistics documentation built on May 5, 2019, 2:41 p.m.