To learn more about EpiX Analytics' work, please visit our modeling applications, white papers, and training schedule.
Page tree

Versions Compared


  • This line was added.
  • This line was removed.
  • Formatting was changed.
Comment: a few additional edits to text

You may have historical data available that you'd like to use to help in forecasting the future. Similarly to fitting parametric distribution to data, it is also possible to fit parametric time-series models to historical data to create forecasts. The main difference is that when fitting a distribution to data, an assumption is that we assumed that they are randomly sampled from independent, identical distributions. In plain words, we assume the data points come from the same distribution and are not relatedcorrelated. For example, if we have measurements of heights of 100 people, each of such measurements is assumed to be the measurement of a randomly selected individual . Timefrom a single population distribution of heights. In contrast, time-series data however assumes that the data is sequentialis by nature sequential as the value in the next period is linked to that of previous periods. For example, we may have weekly height of a person, the monthly the daily price of a commodity , or the yearly sales of a product. Based on this historical data, we would like to estimate what the height, price or sales in the next weeks, months, or years would be, highly dependent on its price in prior days. This type of dependency is called autocorrelation or serial correlation, and must be incorporated in a time-series fit.  

Some important considerations are: