# Summary

The forecast calculation depends entirely on the quality of autocorrelation in the data. If present values are dependent on past data, then the data will give good future predictions for the time series. The Durbin-Watson probe to check the level of autocorrelation of the time series tells us how good the prediction will be by measuring the influence of past data on the current values.

The season component depends on the CMA distance to the data. The season component is determined by the forecast as a factor to move the trend (linear regression) up or down, depending on the cycles of the time series. Comparing the forecast time-series line chart with the original data gives us an idea of how accurate the model's prediction is.

In this chapter, we learned to use the CMA to smooth the peaks and troughs of the seasonal values over the years. The CMA helps to calculate the seasonal trend weight that leads the regression line up and down for the monthly dependable forecasting...