BNEngine-class: Class '"BNEngine"'

In ralmond/EABN: Evidence Accumulation Bayes Net Engine

Class "BNEngine"

Description

A generic engine for handling evidence messages (EvidenceSet objects).

Details

This is the basic class for running the evidence accumulation process. This is actually an abstract class, there are two subclasses: BNEngineMongo, which uses the Mongo database to store student records and as a message queue, and BNEngineNDB, which operates without a database. Note that the BNEngine constructor generates an error.

The following functions form the core of the Engine Protocol:

loadManifest

This loads the network manifest for the PnetWarehouse.

setupDefaultSR

Sets up the default Student Record (used for creating new student records)

configStats

Configures the statistics that are reported in the main loop.

baselineHist

Sets up the baselines for histories.

mainLoop

This runs through a queue of messages, handling the evidence.

handleEvidence

Handles evidence from one scoring context and one user.

accumulateEvidence

Does the actual work of processing the evidence.

getRecordForUser

Fetches the student record for a user, essentially a call to getSR.

logEvidence

Logs the evidence as part of the student record.

updateSM

Updates the student model for the new evidence.

updateStats

Calculates new statistics for the revised student model.

updateHist

Updates the history for the revised student model.

announceStats

Updates other processes about the existance of updated statistics.

Extends

All reference classes extend and inherit methods from "envRefClass".

Methods

app

signature(x = "BNEngine"): Returns the guid identifying the application that this engine is handling.

notifyListeners

signature(sender = "BNEngine"): Notifies other processes that student records have been updated.

fetchNextEvidence

signature(eng = "BNEngine"): Returns the next unprocessed EvidenceSet in the queue.

markProcessed

signature(eng = "BNEngine", eve = "EvidenceSet"): marks the eve argument as processed.

Fields

app:

Object of class character giving an globally unique identifier for the application

srs:

Object of class StudentRecordSet of NULL giving the student record set for the application.

profModel:

Object of class character giving the name of the proficiency model (for the default student record) in the warehouse manifest.

listenerSet:

Object of class ListenerSet giving a set of listeners who will listen for new statistics.

statistics:

Object of class list containing Statistic objects to be run on every update cycle.

histNodes:

Object of class character giving the names of the nodes in the proficiency model whose history will be recorded.

warehouseObj:

Object of class PnetWarehouse which stores the Bayes nets, both evidence models and student models are stored here.

waittime:

Object of class numeric giving the time in seconds the main event loop should wait before checking again for messages.

processN:

Object of class numeric giving the number of times that the main loop should run before stopping. If Inf, then the main loop will run without stopping.

Class-Based Methods

activate():

Sets the flag to indicate that the process is running.

deactivate():

Clears the flag to indicate that the process is no longer running.

shouldHalt():

This function checks the database to see whether or not the flag is set to cause the process to halt after processing the current record..

stopWhenFinished():

This function checks the database to see whether or not the flag is set to cause the process to stop when the event queue is empty.

setHistNodes(nodenames):

Sets the names of the history nodes. Note this should be called before the call to baselineHist or the history nodes will not be set properly in the default student record.

fetchNextEvidence():

Fetches the next evidence set to be handled.

setError(mess, e):

Adds an error flag to an evidence set that generated an error.

getHistNodes():

Retrieves the history nodes.

saveStats(statmat):

Updates the set of statitics associated with this engine.

studentRecords():

Fetches the StudentRecordSet associated with the engine. Note: This method should be called instead of the raw field as it will initialize the field if it is not set up yet.

fetchStats():

Fetches statistic objects from the database.

stats():

Returns the set of Statistic objects associate with the engine.

fetchManifest():

Fetches the network manifest from the database.

setManifest(manifest):

Sets the manifest for the PnetWarehouse.

saveManifest(manifest):

Saves the network manifest to the database.

show():

Providse a printed representation of the database.

setProcessed(mess):

Sets an evidence set message as processed.

warehouse():

Returns the PnetWarehouse associated with this engine. Again, this function should be called in preference to directly accessing the field as it forces initialization when necessary.

evidenceSets():

A reference to the collection of evidence sets.

Author(s)

Russell Almond

References

Almond, Mislevy, Steinberg, Yan and Williamson (2015). Bayesian Networks in Educational Assessment. Springer. Especially Chapter 13.

See Also

Subclasses: BNEngineMongo, BNEngineNDB

Constituent parts: StudentRecordSet, PnetWarehouse

Setup Functions: loadManifest, setupDefaultSR, configStats, baselineHist,

Main Loop Functions: mainLoop, handleEvidence, getRecordForUser, logEvidence, accumulateEvidence, updateRecord, updateSM, updateStats, updateHist, announceStats,

Examples

showClass("BNEngine")

ralmond/EABN documentation built on Aug. 30, 2023, 12:52 p.m.