Vector Joint Model versus Univariate Component Models
Get hourly temperature readings for May 2014 in Champaign, Illinois.
In[1]:= | ![]() X |
In[2]:= | ![]() X |
Use TimeSeriesAggregate to calculate daily minimum and maximum temperatures.
In[3]:= | ![]() X |
Combine them into a vector time series.
In[4]:= | ![]() X |
Out[4]= | ![]() |
Out[5]= | ![]() |
The first part of the data will be used to find a model, while the rest of the data will serve as a reference set for the forecast.
In[6]:= | ![]() X |
The temperatures are cross-correlated.
In[7]:= | ![]() X |
Out[7]= | ![]() |
Fit a vector model to the data.
In[8]:= | ![]() X |
Out[8]= | ![]() |
Forecast the next 5 days.
In[9]:= | ![]() X |
Out[10]= | ![]() |
Find univariate models of the same type but longer orders for both temperatures separately.
In[11]:= | ![]() X |
Out[11]= | ![]() |
In[12]:= | ![]() X |
Out[12]= | ![]() |
Combine univariate forecasts for plotting.
In[13]:= | ![]() X |
Out[13]= | ![]() |
Compare the forecasts.
In[14]:= | ![]() X |
Out[15]= | ![]() |
Plot the vector forecast and the 95% confidence bands.
Out[17]= | ![]() |
Plot the univariate forecasts and the 95% confidence bands.
Out[20]= | ![]() |
Compare both forecasts and the corresponding confidence bands.
Out[21]= | ![]() |