rewindReadapt: Rewind and Readapt given feedback

In mickash/Adaptive-Bayesian-Networks: Adaptive Bayesian Networks.

Description

Given user feedback that a particular variable took a particular value between times t and u this function will: 1. Restore the image closest to but prior to t. 2. Change all data between t and s such that the variable at the specified index takes the specified value. 3. Remove all results of items occurring after the time of the restored image. The rewound' network will be returned. No readaption will have yet occurred. All external inference sessions should be stopped before running this function.

Usage

rewindReadapt(cn, net, from, to, variableIndices, values)

Arguments

cn	An open RODBC connection
net	The network. The network must have a valid image table name in the imageTable field - so network tables must have been initialized.
from	The time from which the data alteration is to occur. This should be a valid date-time string. For MySQL this is YYYY-MM-DD HH:MM:SS.
to	The time to which the data alteration is to occur. This should be a valid date-time string. For MySQL this is YYYY-MM-DD HH:MM:SS.
variableIndices	The indices of the variables whose values are known in the specified interval.
values	The values for the above variables.

Value

The network set to the state of the appropriate prior image. No readaption will have yet occurred.

mickash/Adaptive-Bayesian-Networks documentation built on Nov. 14, 2019, 12:14 a.m.