For which value do you believe the model gives a better forecast? Explain.

Wal Mart Revenue

=

-13.80 +0.000092Retail Sales Index

For a Retail Sales Index(RSI) 350000 the prediction is 18.3466 or $18.3466 billion. The Standard error is .426868. The CI is between 17.4817 and 19.2115.

For a RSI 500000 the prediction is 32.1254 or $32.1254 billion. The standard error is 2.1164. The CI is between 27.8448 and 36,4060.

The smaller the standard error, the better the regression model fits the data set. The 350000 prediction is more accurate. In terms of value for Walmart, the model of RSI 350000 implies that the prediction will be $18.3466 billion, and if it's off it will not be off by more than 2 standard deviations or approximately $852 million. So, on the low side, the actual data point maybe $17.4817 billion and the upper side is $19.2115. From a forecasting perspective, Walmart could plan business activities within this spread and be confident it will be correct.

The RSI 500000 is less accurate because the SE is 2.1164. Therefore, the model could be off by 2 standard deviations in either direction. Each standard deviation is approximately $2.1 billion. Therefore, the model could be off by $4.2 billion in either direction. Personally, I wouldn't want to have to explain why the actual data point missed the mark by $ billions.

Seeing a comparison on CI versus PI is interesting. CI tries to quantify the variability of an estimate through many data points. PI tries to capture the variability of a future data point. PI is typically more comprehensive.

Minitab throws up a XX marker in the calculation, implying that there are highly unusual levels in the RSI 500000 PI model. Most likely because the prediction is significantly higher than anything in the data set. It reminds me of a plane taking off. Everything is larger and visible on the ground but as the plane climbs objects on the ground begin to shrink.