createTrainingSpecs: Run Training Functions
In KNewhart/ADPCA: ADPCA for Process Monitoring of WWTP

Using only the data from the start to the day before testing day, train the ADPCA model using functions from mvMonitoring

1 2	createTrainingSpecs(data, testingDay, rollingWindowDays, alpha = 0.01, faultsToTriggerAlarm = 5, statistic = "T2", by = "Time")

`data`	Dataframe of process data
`testingDay`	Date object, all dates before this date will be included in the training analysis
`rollingWindowDays`	Number of observations used in the rolling window
`alpha`	Accuracy of analysis, defaults to 0.01
`faultsToTriggerAlarm`	Number of abnormal observations to be considered an alarm, defaults to five
`statistic`	Use "T2" or "SPE"
`by`	"Time" or number of observations to use in training window

trainingSpecHolder <- createTrainingSpecs(data = dataBR[[i]],
testingDay = testingDay,
trainObs = rollingWindowObs,
updateFreq = updateObs,
alpha = alphaN,
faultsToTriggerAlarm = faultsToTriggerAlarm)