fitModels: Fits the HMM model for each device.

In bogdanoancea/deduplication: R package for computing the duplicity probility for each mobile device.

Description

Fits the HMM model for each device using the fit() function from destim package. The computations are done in parallel to reduce the running time using all the available cores. This function creates a cluster of working nodes, splits the devices equally among the working nodes and assigns a partition of devices to each node in the cluster. For Unix-like operating systems, this functions uses a "FORK" cluster while for Windows it uses a "SOCK" cluster.

Usage

fitModels(ndevices, model, connections)

Arguments

ndevices	The number of devices.
model	The HMM model returned by getGenericModel() function. This model is fitted for each device.
connections	A matrix whose elements are the antenna ID where a device is connected at every time instant. This matrix is returned by getConnections() function.

Value

A vector of log likelihoods computed using the fitted model for each device.

bogdanoancea/deduplication documentation built on Dec. 2, 2020, 11:22 p.m.