def testVectorAutoRegressiveMethodMetrics():
numberOfObservation = 12
#reading the full dataset
vectorAutoRegressiveMethodDataset = importVectorAutoRegressiveMethodDataset(
"M2SLMoneyStock.csv", "PCEPersonalSpending.csv")
#reading the forecasted values
vectorAutoRegressiveMethodForecastedValues = readVectorAutoRegressiveMethodForecastedValues(
)
#rmse for money
rmseForMoneyForecasting = rmse(
vectorAutoRegressiveMethodDataset["Money"][-numberOfObservation:],
vectorAutoRegressiveMethodForecastedValues["MoneyForecast"])
#rmse for spending
rmseForSpendingForecasting = rmse(
vectorAutoRegressiveMethodDataset["Spending"][-numberOfObservation:],
vectorAutoRegressiveMethodForecastedValues["SpendingForecast"])
print(rmseForMoneyForecasting) #43.71049653558938
print(rmseForSpendingForecasting) #37.00117516940808
def plotVectorAutoRegressiveMethodPredictedValues():
#reading the forecasted values
vectorAutoRegressiveMethodForecastedValues = readVectorAutoRegressiveMethodForecastedValues(
)
#reading the full dataset
vectorAutoRegressiveMethodDataset = importVectorAutoRegressiveMethodDataset(
"M2SLMoneyStock.csv", "PCEPersonalSpending.csv")
visualizeVectorAutoRegressiveMethodPredictedValuesForMoney(
vectorAutoRegressiveMethodDataset,
vectorAutoRegressiveMethodForecastedValues)
visualizeVectorAutoRegressiveMethodPredictedValuesForSpending(
vectorAutoRegressiveMethodDataset,
vectorAutoRegressiveMethodForecastedValues)
def trainVectorAutoRegressiveMethodModelOnFullDataset():
vectorAutoRegressiveMethodDataset = importVectorAutoRegressiveMethodDataset(
"M2SLMoneyStock.csv", "PCEPersonalSpending.csv")
#training model on the whole dataset
vectorAutoRegressiveMethodModel = VAR(vectorAutoRegressiveMethodDataset)
#we are taking p = 5 as we have created different models based on the different p values.
#Model gives minimum aic and bic for p =5
vectorAutoRegressiveMethodModelResult = vectorAutoRegressiveMethodModel.fit(
5)
#saving the model in pickle files
saveVectorAutoRegressiveMethodModelForFullDataset(
vectorAutoRegressiveMethodModelResult)
print(vectorAutoRegressiveMethodModelResult.summary())
def preprocess():
vectorAutoRegressiveMethodDataset = importVectorAutoRegressiveMethodDataset(
"M2SLMoneyStock.csv", "PCEPersonalSpending.csv")
#taking the first difference
vectorAutoRegressiveMethodDatasetFirstDiff = vectorAutoRegressiveMethodDataset.diff(
)
#taking the second difference. Second difference data is stationary
vectorAutoRegressiveMethodDatasetSecondDiff = vectorAutoRegressiveMethodDatasetFirstDiff.diff(
)
#dropping missing values
vectorAutoRegressiveMethodDatasetSecondDiff = vectorAutoRegressiveMethodDatasetSecondDiff.dropna(
)
X_train, X_test = splitVectorAutoRegressiveMethodDataset(
vectorAutoRegressiveMethodDatasetSecondDiff)
saveTrainingAndTestingDataset(X_train, X_test)
def testVectorAutoRegressiveMethodModel():
#reading the full dataset
vectorAutoRegressiveMethodDataset = importVectorAutoRegressiveMethodDataset(
"M2SLMoneyStock.csv", "PCEPersonalSpending.csv")
#reading testing data
X_train = readVectorAutoRegressiveMethodXTrain()
#reading model from pickle file
vectorAutoRegressiveMethodModel = readVectorAutoRegressiveMethodModel()
#Unlike the VARMAX model we'll use in upcoming sections, the VAR .forecast() function
#requires that we pass in a lag order number of previous observations as well.
#Unfortunately this forecast tool doesn't provide a DateTime index - we'll have to do that manually.
#forecast for next 12 months
predictedValues = vectorAutoRegressiveMethodModel.forecast(
y=X_train.values[-5:], steps=12)
idx = pd.date_range('1/1/2015', periods=12, freq='MS')
vectorAutoRegressiveMethodForecastedValues = pd.DataFrame(
predictedValues, index=idx, columns=['Money2d', 'Spending2d'])
numberOfObsevation = 12
#Invert the transformation
#Remember that the forecasted values represent second-order differences.
#To compare them to the original data we have to roll back each difference.
#To roll back a first-order difference we take the most recent value on the training side of the original series,
#and add it to a cumulative sum of forecasted values.
#When working with second-order differences we first must perform this operation on the most recent first-order difference.
# Add the most recent first difference from the training side of the original dataset to the forecast cumulative sum
vectorAutoRegressiveMethodForecastedValues['Money1d'] = (
vectorAutoRegressiveMethodDataset['Money'].iloc[-numberOfObsevation -
1] -
vectorAutoRegressiveMethodDataset['Money'].iloc[-numberOfObsevation -
2]
) + vectorAutoRegressiveMethodForecastedValues['Money2d'].cumsum()
# Now build the forecast values from the first difference set
vectorAutoRegressiveMethodForecastedValues[
'MoneyForecast'] = vectorAutoRegressiveMethodDataset['Money'].iloc[
-numberOfObsevation -
1] + vectorAutoRegressiveMethodForecastedValues['Money1d'].cumsum(
)
# Add the most recent first difference from the training side of the original dataset to the forecast cumulative sum
vectorAutoRegressiveMethodForecastedValues['Spending1d'] = (
vectorAutoRegressiveMethodDataset['Spending'].iloc[-numberOfObsevation
- 1] -
vectorAutoRegressiveMethodDataset['Spending'].iloc[-numberOfObsevation
- 2]
) + vectorAutoRegressiveMethodForecastedValues['Spending2d'].cumsum()
# Now build the forecast values from the first difference set
vectorAutoRegressiveMethodForecastedValues[
'SpendingForecast'] = vectorAutoRegressiveMethodDataset[
'Spending'].iloc[-numberOfObsevation -
1] + vectorAutoRegressiveMethodForecastedValues[
'Spending1d'].cumsum()
#saving the foreasted values without lag
saveVectorAutoRegressiveMethodForecastedValues(
vectorAutoRegressiveMethodForecastedValues)
def testIsDatasetStationary():
vectorAutoRegressiveMethodDataset = importVectorAutoRegressiveMethodDataset(
"M2SLMoneyStock.csv", "PCEPersonalSpending.csv")
#agumentedDickeyFullerTest(vectorAutoRegressiveMethodDataset["Money"])
# =============================================================================
# Augmented Dickey-Fuller Test:
# ADF test statistic 4.239022
# p-value 1.000000
# # lags used 4.000000
# # observations 247.000000
# critical value (1%) -3.457105
# critical value (5%) -2.873314
# critical value (10%) -2.573044
# Weak evidence against the null hypothesis
# Fail to reject the null hypothesis
# Data has a unit root and is non-stationary
# =============================================================================
#agumentedDickeyFullerTest(vectorAutoRegressiveMethodDataset["Spending"])
# =============================================================================
# Augmented Dickey-Fuller Test:
# ADF test statistic 0.149796
# p-value 0.969301
# # lags used 3.000000
# # observations 248.000000
# critical value (1%) -3.456996
# critical value (5%) -2.873266
# critical value (10%) -2.573019
# Weak evidence against the null hypothesis
# Fail to reject the null hypothesis
# Data has a unit root and is non-stationary
# =============================================================================
vectorAutoRegressiveMethodDatasetFirstDiff = vectorAutoRegressiveMethodDataset.diff(
)
#agumentedDickeyFullerTest(vectorAutoRegressiveMethodDatasetFirstDiff["Money"], title = "Money First Differnce")
# =============================================================================
# Augmented Dickey-Fuller Test: Money First Differnce
# ADF test statistic -2.057404
# p-value 0.261984
# # lags used 15.000000
# # observations 235.000000
# critical value (1%) -3.458487
# critical value (5%) -2.873919
# critical value (10%) -2.573367
# Weak evidence against the null hypothesis
# Fail to reject the null hypothesis
# Data has a unit root and is non-stationary
# =============================================================================
#agumentedDickeyFullerTest(vectorAutoRegressiveMethodDatasetFirstDiff["Spending"], title = "Spending First Differnce")
# =============================================================================
# Augmented Dickey-Fuller Test: Spending First Differnce
# ADF test statistic -7.226974e+00
# p-value 2.041027e-10
# # lags used 2.000000e+00
# # observations 2.480000e+02
# critical value (1%) -3.456996e+00
# critical value (5%) -2.873266e+00
# critical value (10%) -2.573019e+00
# Strong evidence against the null hypothesis
# Reject the null hypothesis
# Data has no unit root and is stationary
# =============================================================================
vectorAutoRegressiveMethodDatasetSecondDiff = vectorAutoRegressiveMethodDatasetFirstDiff.diff(
)
agumentedDickeyFullerTest(
vectorAutoRegressiveMethodDatasetSecondDiff["Money"],
title="Money Second Differnce")
# =============================================================================
# Augmented Dickey-Fuller Test: Money Second Differnce
# ADF test statistic -7.077471e+00
# p-value 4.760675e-10
# # lags used 1.400000e+01
# # observations 2.350000e+02
# critical value (1%) -3.458487e+00
# critical value (5%) -2.873919e+00
# critical value (10%) -2.573367e+00
# Strong evidence against the null hypothesis
# Reject the null hypothesis
# Data has no unit root and is stationary
# ============================================================================
agumentedDickeyFullerTest(
vectorAutoRegressiveMethodDatasetSecondDiff["Spending"],
title="Spending Second Differnce")
def plotTheSourceData():
vectorAutoRegressiveMethodDataset = importVectorAutoRegressiveMethodDataset(
"M2SLMoneyStock.csv", "PCEPersonalSpending.csv")
visualizeSourceDataPlot(vectorAutoRegressiveMethodDataset)